Navigating a nonholonomic robot in a cluttered, unknown environment requires accurate perception and precise motion control for real-time collision avoidance. This article presents neural proximal alternating-minimization network (NeuPAN): a real-time, highly accurate, map-free, easy-to-deploy, and environment-invariant robot motion planner. Leveraging a tightly coupled perception-to-control framework, NeuPAN has two key innovations compared to existing approaches: first, it directly maps raw point cloud data to a latent distance feature space for collision-free motion generation, avoiding error propagation from the perception to control pipeline; second, it is interpretable from an end-to-end model-based learning perspective. The crux of NeuPAN is solving an end-to-end mathematical model with numerous point-level constraints using a plug-and-play proximal alternating-minimization network, incorporating neurons in the loop. This allows NeuPAN to generate real-time, physically interpretable motions. It seamlessly integrates data and knowledge engines, and its network parameters can be fine-tuned via back propagation. We evaluate NeuPAN on a ground mobile robot, a wheel-legged robot, and an autonomous vehicle, in extensive simulated and real-world environments. Results demonstrate that NeuPAN outperforms existing baselines in terms of accuracy, efficiency, robustness, and generalization capabilities across various environments, including the cluttered sandbox, office, corridor, and parking lot. We show that NeuPAN works well in unknown and unstructured environments with arbitrarily shaped objects, transforming impassable paths into passable ones.

The substantial memory demands of model weights and key-value (KV) caches often lead to severe memory bottlenecks in LLM serving. Existing systems address this by offloading KV caches to host memory and rapidly restoring them on demand before decoding. However, these approaches are too coarse-grained and fail to fully exploit the combined computational and storage capabilities of GPUs.

Serving Large Language Models (LLMs) in production faces significant challenges from highly variable request patterns and severe resource fragmentation in serverless clusters. Current systems rely on static pipeline configurations that struggle to adapt to dynamic workload conditions, leading to substantial inefficiencies.

While spot instances offer a cost-effective alternative to on-demand cloud resources, they introduce reliability challenges for latency-sensitive microservices due to preemption risks and unpredictable provisioning delays. Conventional resource management systems, which often rely on assumptions of immediate instance availability, fail to account for these operational realities—resulting in increased risk of SLO violations when deployed in spot-based environments.

Transformer-based models have enabled unprecedented capabilities across language, vision, and multimodal tasks. On-device fine-tuning of transformer models offers a privacy-preserving path to personalized AI, yet remains inefficient on mobile GPUs due to severe memory constraints and frequent layout transformations in attention mechanism during training. Existing mobile training frameworks either use unified layouts for forward and backward passes - leading to fragmented memory access and poor GPU utilization during backpropagation - or rely on explicit layout conversions, which introduce significant transformation overhead.To overcome this, we propose FBLayout, a layout-aware framework that co-designs tensor organization with mobile GPU platforms. FBLayout introduces: (1) a unified R-Tile layout for multidimensional reductions across forward/backward passes; (2) tile-based index transformation to eliminate physical data movement; and (3) activation-guided layout selection to propagate efficient layouts globally. Evaluations on seven transformer models across different mobile phones (including ARM Mali and Qualcomm Adreno GPUs) show that FBLayout achieves 2.2-5.7× speedup over MNN, TFLite, and TVM, while significantly improving cache efficiency and reducing memory footprint, enabling practical on-device large model fine-tuning.

Modern LLM serving systems increasingly adopt elastic inference pipelines where stages frequently join, leave, and migrate across nodes. However, existing GPU communication frameworks like NCCL assume static topologies, causing routing failures and P99 latency spikes during worker transitions that violate sub-millisecond tail latency requirements. We present Connex, a communication system that elevates endpoint mobility from exceptional failure to first-class primitive. Rather than optimizing individual mechanisms in isolation, Connex defines a mobility contract that the communication layer enforces whenever workers join, leave, or migrate while token streams, activations, or KV transfers are in flight. The contract is realized through three cooperating mechanisms: (1) epoch-based routing that bounds staleness without global coordination, (2) explicit handover protocols that preserve stream ordering and provide exactly-once delivery across migrations, and (3) credit-based backpressure with traffic-class isolation that prevents churn-induced interference with latency-critical paths. Evaluation on a 5-node GPU cluster under synthetic and production-derived churn shows that Connex reduces P99 tail spikes by up to 85% compared to NCCL-based baselines, achieves sub-second cutover, and maintains 100% goodput at moderate loads where baselines collapse to 0–28%, while incurring less than 5% steady-state overhead.

ABSTRACT Objective Large Language Models (LLMs) are increasingly deployed in modern AI infrastructure, creating a strong demand for high‐throughput and resource‐efficient serving systems. Disaggregated LLM serving, which decouples prompt prefill from auto‐regressive decode to accommodate their heterogeneous compute and memory characteristics, has emerged as a promising architecture. However, existing disaggregated serving systems suffer from three fundamental limitations: static resource allocation that fails to adapt to highly dynamic workloads, severe load imbalance between compute‐bound prefill and memory‐bound decode stages, and prefix‐cache‐aware routing that skews load distribution and creates performance hotspots. These issues collectively limit resource utilization, scalability, and the ability to meet service level objectives (SLOs) under real‐world workloads. Methods To address these challenges, we propose BanaServe, a dynamic orchestration framework for disaggregated LLM serving that continuously rebalances both computational and memory resources across prefill and decode instances. BanaServe introduces three key mechanisms: (i) layer‐level weight migration to enable coarse‐grained redistribution of computation, (ii) attention‐level Key–Value (KV) cache migration for fine‐grained memory load balancing, and (iii) a Global KV Cache Store with layer‐wise overlapped transmission to decouple routing decisions from cache placement. Together, these mechanisms eliminate cache‐induced hotspots and allow routers to perform purely load‐aware scheduling with minimal latency overhead. BanaServe is implemented on top of state‐of‐the‐art LLM serving frameworks, including vLLM and DistServe. Results We evaluate BanaServe under diverse and challenging workloads, including long‐context inference, bursty request arrivals, and mixed prompt–generation patterns. Experimental results show that, compared to vLLM, BanaServe improves throughput by 1.2–3.9× and reduces total processing time by 3.9%–78.4%. In comparison with DistServe, BanaServe achieves 1.1–2.8× higher throughput while reducing latency by 1.4%–70.1%. These gains are consistent across workload variations, demonstrating BanaServe's robustness under highly dynamic serving conditions. Conclusion BanaServe demonstrates that dynamic, multi‐granularity resource rebalancing and cache‐decoupled routing are essential for efficient disaggregated LLM serving. By jointly addressing resource elasticity, stage imbalance, and cache‐induced load skew, BanaServe substantially improves throughput, latency, and resource utilization in real‐world deployments. This work provides a practical and scalable foundation for next‐generation LLM serving systems operating under dynamic and heterogeneous workloads.

ABSTRACT Objective Microservices transform traditional monolithic applications into lightweight, loosely coupled application components and have been widely adopted in many enterprises. Cloud platform infrastructure providers enhance the resource utilization efficiency of microservices systems by co‐locating different microservices. However, this approach also introduces resource competition and interference among microservices. Designing interference‐aware strategies for large‐scale, co‐located microservice clusters is crucial for enhancing resource utilization and mitigating competition‐induced interference. These challenges are further exacerbated by unreliable metrics, application diversity, and node heterogeneity. Methods In this paper, we first analyze the characteristics of large‐scale and co‐located microservices clusters at Alibaba and further discuss why cycle per instruction (CPI) is adopted as a metric for interference measurement in large‐scale production clusters, as well as how to achieve accurate prediction of CPI through multi‐dimensional metrics. Based on CPI interference prediction and analysis, we also present the design of the C‐Koordinator platform, an open‐source solution utilized in Alibaba cluster, which incorporates co‐location and interference mitigation strategies. Results The interference prediction models consistently achieve over 90.3% accuracy, enabling precise prediction and rapid mitigation of interference in operational environments. As a result, application latency is reduced and stabilized across all percentiles (P50, P90, P99) response time (RT), achieving improvements ranging from 16.7% to 36.1% under various system loads compared with state‐of‐the‐art system. Conclusion These results demonstrate the system's ability to maintain smooth application performance in co‐located environments.

In recent years, the Mixture-of-Experts (MoE) architecture has been widely applied to large language models (LLMs), providing a promising solution that activates only a subset of the model’s parameters during computation, thereby reducing overall memory requirements and allowing for faster inference compared to dense models. Despite these advantages, existing systems still face issues of low efficiency due to static model placement and lack of dynamic workloads adaptation. This leads to suboptimal resource utilization and increased latency, especially during bursty requests periods.To address these challenges, this paper introduces Brownout-Serve, a novel serving framework designed to optimize inference efficiency and maintain service reliability for MoE-based LLMs under dynamic computational demands and traffic conditions. BrownoutServe introduces “united experts” that integrate knowledge from multiple experts, reducing the times of expert access and inference latency. Additionally, it proposes a dynamic brownout mechanism to adaptively adjust the processing of certain tokens, optimizing inference performance while guaranteeing service level objectives (SLOs) are met. Our evaluations show the effectiveness of BrownoutServe under various workloads: it achieves up to 2.46× throughput improvement compared to state-of- the-art systems and reduces SLO violations by up to 90.28%, showcasing its robustness under bursty traffic while maintaining acceptable inference accuracy.

Depression is one of the most common mood disorders and the number of patients increases significantly in recent years. Due to the lack of biomarkers, conversation between patients and psychiatrists is still the main clinical diagnostic method which is easily influenced by subjectivity of both patients and psychiatrists. Synthetic House-tree-person test (S-HTP), a convenient and efficient mental assessment tool, minimizes subjective influences from patients, while its effectiveness is limited by the professional ability of analyst. Here we introduce a deep learning model DeHTP, a flexible and convenient depression detection method based on S-HTP without interaction between people. Experimental results demonstrate that DeHTP achieves 0.963 AUC and 0.9 accuracy, and outperforms the conventional manual analysis of S-HTP, which is conducted on the guideline of 50 conclusions from previous study related to depression. In addition, it reveals 22 depression-correlated drawing features aligned with conclusions above from the perspective of our proposed model. Leveraging the advantages of deep learning and S-HTP, this approach has the potential for widespread promotion and adoption as the available tool for daily self-mental monitoring, as well as the promising auxiliary diagnostic method in clinical.

Open source software supply-chain attacks, once successful, can exact heavy costs in mission-critical applications. As open source ecosystems for deep learning flourish and become increasingly universal, they present attackers previously unexplored avenues to code-inject malicious backdoors in deep neural network models. This article proposes Flareon , a small, stealthy, seemingly harmless code modification that specifically targets the data augmentation pipeline with motion-based triggers. Flareon neither alters ground-truth labels, nor modifies the training loss objective, nor does it assume prior knowledge of the victim model architecture, training data, and training hyperparameters. Yet, it has a surprisingly large ramification on training—models trained under Flareon learn powerful target-conditioned (or “ all2all ”) backdoors. We also proposed a learnable variant of Flareon that is even stealthier in terms of added perturbations. The resulting models can exhibit high attack success rates for any target choices and better clean accuracies than backdoor attacks that not only seize greater control but also assume more restrictive attack capabilities. We also demonstrate the resilience of Flareon against a wide range of defenses. Flareon is fully open source and available online to the deep learning community.

Traditional crowd sensing based indoor localization methods rely on large scale pre-collected fingerprint data to construct a radio map with cumbersome prior preparation. However, when they lack floor plan information or only have a little of data is willing to share, the tracking accuracy degrades significantly. In this paper, we propose a singular value decomposition (SVD) track matching scheme to obtain an effective radio map based on small scale crowd sensing data, which is a non-learning based system (SVD-CSP). SVD-CSP fuses received signal strength indicator (RSSI), inertial measurement unit (IMU), and magnetic field strength to label surrounding WiFi access points as marker points. The proposed scheme uses SVD method to directly compute the rotation matrix and displacement vector among the crowd sensing trajectories and attain the reliable tracks. The radio map is constructed and users are tracked according to our developed bidirectional Bayesian filter, which contains forward filter and reverse filter. The density-based spatial clustering of applications with noise (DBSCAN) is embedded within the forward filter to improve the radio map quality. Meanwhile, the reverse filter fuses pedestrian dead reckoning (PDR) and radio map-based localization to track users. Experimental results demonstrate that SVD-CSP can achieve robust localization using extremely sparse crowd trajectories (e.g., 4 trajectories in a 648 m2scenario, 30 trajectories in a 2856 m2scenario) without deep learning training or infrastructure knowledge.

Hybrid reconfigurable intelligent surfaces (HRIS) can enable 6 G vehicle-to-everything (V2X) system to attain promising localization and communication performance due to its flexible beamforming feature. However, without jointly designing the HRIS control and system resource allocation scheme, the HRIS-V2X system can not adapt to the dynamic environment efficiently. In addition, the optimization of HRIS reflectivity and communication time slice are both non convex and nonlinear problems. In this paper, we propose an asynchronous time division multiplexing (ATDM) protocol for the HRIS-V2X system to meet integrated localization and communications requirements. We analyze the role of HRIS in signal transmission according to squared position error bound (SPEB) and achievable rate (AR). Then, we propose an adaptive block coordinate descent (ABCD) algorithm to optimize the localization accuracy and channel transmission capability, which includes two parts: the time optimization and the reflectivity optimization. Time optimization employs the iterative projection method to find the optimal time slice scheme satisfying AR constraints. Reflectivity optimization uses the Adagrad method with an adaptive learning rate to gradually achieve the optimal reflectivity scheme. The simulation results indicate that our proposed ABCD algorithm has achieved a maximum 94.1% reduction in SPEB compared to greedy algorithm, genetic algorithm (GA), artificial rabbits optimization (ARO) and particle swarm optimization (PSO).

In contemporary cloud architectures, an increasing number of cloud service providers are adopting programmable switches to deliver cloud network services (e.g., load-balancing gateways) for millions of tenants. The rate limiters are essential for cloud networks to execute network policies such as congestion control and traffic isolation on programmable switches. Most existing rate limiters utilize the token bucket algorithm, which suffers from scalability issues and substantial control overhead, impacting bandwidth utilization. The WF2Q+ (Worst-case FairWeighted Fair Queueing Plus) algorithm, known for its scalability and accuracy, is gaining traction but faces challenges on programmable switches. The hardware limitations hinder key operations of WF2Q+ (sorting and scheduling) in a single switch pipeline. This paper introduces DRAGONKING, a novel rate-limiting system that enables WF2Q+ through a multi-pipeline sorting and scheduling design. DRAGONKING enhances scalability and throughput while maintaining high accuracy. It achieves this by strategically balancing bandwidth across multiple pipelines, allowing for timely packet scheduling. DRAGONKING is implemented and evaluated on Barefoot Tofino switch. Results show that DRAGONKING supports up to two million entries and delivers line-rate throughput, achieving 1.9× improvements compared with token bucket-based limiters. Moreover, it can maintain over 99% accuracy, precisely enforcing rate limits from 10 Gbps to 100 Gbps.

Large language models increasingly rely on pipeline parallelism for distributed inference, but existing systems face critical challenges in serverless environments: heterogeneous request distributions across pipeline stages and unpredictable workload patterns requiring rapid elasticity. Traditional static resource allocation fails to address pipeline-specific bottlenecks and cold start delays inherent in serverless architectures. We propose QUART, a workload-adapted resource allocation system for LLM distributed serving in serverless clusters. QUART introduces pipeline-aware resource management through: (1) latency-aware critical stage identification using coefficient of variation (CV)-based burst propagation analysis, (2) dynamic replica allocation with proportional-integral-derivative (PID) control for congested stages, and (3) hierarchical parameter caching with copy-on-write mechanisms enabling sub-second serverless scaling. The system addresses serverless-specific challenges through cache-aware scheduling that maintains model parameters in memory, eliminating disk I/O overhead during rapid scaling events. Evaluation with real-world workloads shows QUART reduces average response latency by up to 87.1% compared to existing serverless inference systems while achieving 2.37x improvement in goodput.

Deploying large language models (LLMs) in realtime systems is challenging due to their high resource demands and privacy concerns. We propose Floe a hybrid federated learning framework designed for latency-sensitive, resourceconstrained environments. Floeombines a cloud-based blackbox LLM with lightweight small language models (SLMs) on edge devices to enable low-latency, privacy-preserving inference. Personal data and fine-tuning remain on-device, while the cloud LLM contributes general knowledge without exposing proprietary weights. A heterogeneity-aware LoRA adaptation strategy ensures efficient edge deployment across diverse hardware, and a logit-level fusion mechanism enables real-time coordination between edge and cloud models. Experiments demonstrate that Floenhances user privacy and personalization, while significantly improving model performance and reducing inference latency on edge devices under real-time constraints, compared to baseline approaches.

To meet strict Service-Level Objectives (SLO), contemporary Large Language Models (LLMs) decouple the prefill and decoding stages and place them on separate GPUs to mitigate the distinct bottlenecks inherent to each stage. However, the heterogeneity of LLM workloads causes producer-consumer imbalance between the two instance types in such disaggregated architecture. To address this problem, we propose DOPD (Dynamic Optimal Prefill/Decoding), a dynamic LLM inference system that adjusts instance allocations to achieve an optimal prefill-to-decoding (P/D) ratio based on real-time load monitoring. Combined with an appropriate request-scheduling policy, DOPD effectively resolves imbalances between prefill and decoding instances and mitigates resource allocation mismatches due to mixed-length requests under high concurrency. Experimental evaluations show that, compared with vLLM and DistServe (representative aggregation-based and disaggregation-based approaches), DOPD improves overall system goodput by up to$1.5\times$, decreases P90 time-to-first-token (TTFT) by up to 67.5%, and decreases P90 time-per-output-token (TPOT) by up to 22.8%. Furthermore, our dynamic P/D adjustment technique performs proactive reconfiguration based on historical load, achieving over 99% SLO attainment while using fewer additional resources.

There are two crucial aspects of reliable autonomous driving systems: the reasoning behind decision-making and the precision of environmental perception. This paper introduces DME-Driver, a new autonomous driving system that enhances performance and robustness by fully leveraging the two crucial aspects. This system comprises two main models. The first, the Decision Maker, is responsible for providing logical driving instructions. The second, the Executor, receives these instructions and generates precise control signals for the vehicles. To ensure explainable and reliable driving decisions, we build the Decision-Maker based on a large vision language model. This model follows the logic employed by experienced human drivers and simulates making decisions in a safe and reasonable manner. On the other hand, the generation of accurate control signals relies on precise and detailed environmental perception, where 3D scene perception models excel. Therefore, a planning-oriented perception model is employed as the Executor. It translates the logical decisions made by the Decision-Maker into accurate control signals for the self-driving cars. To effectively train the proposed system, a new dataset named Human-driver Behavior and Decision-making (HBD) dataset has been collected. This dataset encompasses a diverse range of human driver behaviors and their underlying motivations. By leveraging this dataset, our system achieves high-precision planning accuracy through a logical thinking process.

Federated Learning (FL) enables collaborative learning from distributed data while preserving the privacy of participating clients. While supervised federated learning with labeled data has made notable strides and achieved success, federated semi-supervised learning (FSSL) lags in its progress. Existing works for FSSL heavily rely on fully-labeled clients, while ignoring the distribution of pseudo-labels generated from skewed unlabeled data. In this work, we offer empirical and theoretical insights into the challenges encountered when applying conventional semi-supervised algorithms in the federated regime. Specifically, we highlight how the inherent data heterogeneity in FSSL can exacerbate issues within the pseudo-labeling process. Motivated by these observations, we propose federated learning with progressive distribution matching (FedPDM) to regularize the distribution of pseudo-labels, aiming to progressively reshape it to align with the ground-truth distribution. The matching problem could be formulated as an optimal transport (OT) problem and efficiently solved by Sinkhorn-Knopp iteration. Through extensive experiments, we demonstrate the superiority of FedPDM on a variety of models and datasets compared with prior arts for FSSL.

Accurate trajectory prediction is essential for the safety and efficiency of autonomous driving. Traditional models often struggle with real-time processing, capturing non-linearity and uncertainty in traffic environments, efficiency in dense traffic, and modeling temporal dynamics of interactions. We introduce NEST (Neuromodulated Small-world Hypergraph Trajectory Prediction), a novel framework that integrates Small-world Networks and hypergraphs for superior interaction modeling and prediction accuracy. This integration enables the capture of both local and extended vehicle interactions, while the Neuromodulator component adapts dynamically to changing traffic conditions. We validate the NEST model on several real-world datasets, including nuScenes, MoCAD, and HighD. The results consistently demonstrate that NEST outperforms existing methods in various traffic scenarios, showcasing its exceptional generalization capability, efficiency, and temporal foresight. Our comprehensive evaluation illustrates that NEST significantly improves the reliability and operational efficiency of autonomous driving systems, making it a robust solution for trajectory prediction in complex traffic environments.

A new trend in the computer vision community is to capture objects of interest following flexible human command represented by a natural language prompt. However, the progress of using language prompts in driving scenarios is stuck in a bottleneck due to the scarcity of paired prompt-instance data. To address this challenge, we propose the first object-centric language prompt set for driving scenes within 3D, multi-view, and multi-frame space, named NuPrompt. It expands nuScenes dataset by constructing a total of 40,147 language descriptions, each referring to an average of 7.4 object tracklets. Based on the object-text pairs from the new benchmark, we formulate a novel prompt-based driving task, \ie, employing a language prompt to predict the described object trajectory across views and frames. Furthermore, we provide a simple end-to-end baseline model based on Transformer, named PromptTrack. Experiments show that our PromptTrack achieves impressive performance on NuPrompt. We hope this work can provide some new insights for the self-driving community.

To enhance autonomous driving, innovative approaches have been proposed to generate simulated LiDAR data. However, these methods often face challenges in producing high-quality and controllable foreground objects. To cater to the needs of object-aware tasks in 3D perception, we introduce OLiDM, a novel framework capable of generating controllable and high-fidelity LiDAR data at both the object and scene levels. OLiDM consists of two pivotal components: the Object-Scene Progressive Generation (OPG) module and the Object Semantic Alignment (OSA) module. OPG adapts to user-specific prompts to generate desired foreground objects, which are subsequently employed as conditions in scene generation, ensuring controllable and diverse output at both the object and scene levels. This also facilitates the association of user-defined object-level annotations with the generated LiDAR scenes. Moreover, OSA aims to rectify the misalignment between foreground objects and background scenes, enhancing the overall quality of the generated objects. The broad efficacy of OLiDM is demonstrated across both unconditional and conditional LiDAR generation tasks, as well as 3D perception tasks. Specifically, on the KITTI-360 dataset, OLiDM surpasses prior state-of-the-art methods such as UltraLiDAR by 11.8 in FPD, producing data that closely mirrors real-world distributions. Additionally, in sparse-to-dense LiDAR completion, OLiDM achieves a significant improvement over LiDARGen, with a 57.47% increase in semantic IoU. Moreover, in 3D object detection, OLiDM enhances the performance of mainstream detectors by 2.4% in mAP and 1.9% in NDS, underscoring its potential in advancing 3D perception models.

In a unified resource scheduling architecture, containers within the same microservice often encounter temporal and spatial performance imbalance when deployed in large-scale shared clusters. As a result, the commonly employed load-balancing approach often leads to substantial resource wastage as applications are frequently over-provisioned to meet service level agreements (SLAs).

This paper presents a comprehensive analysis of diffusion model serving challenges in production cloud environments. We examine the unique computational patterns and resource requirements that distinguish diffusion model serving from traditional ML workloads, revealing fundamental systemlevel challenges from their multi-stage pipeline architectures. Our analysis is based on a dataset collected from a commercial image generation service processing 3.5 million requests across 300+ GPUs of production operation.

Advances in large language models (LLMs) have opened up new possibilities across various fields, fueling a new wave of interactive AI applications such as DeepSeek and ChatGPT. Inference serving systems play a crucial role in supporting these applications. Recent research indicates that when crossserver parallelization is enabled in inference serving systems, data synchronization overhead can exceed 65% of the total inference delay, making the reduction of communication overhead essential for speeding up inference. While existing systems accelerate cross-server communications by offloading synchronization operations to programmable switches, they often suffer from limited aggregation throughput under bursty traffic conditions, posing challenges for homogeneous network environments. To address these challenges, we propose HeroServe, an innovative inference serving system that leverages heterogeneous networks to accelerate data synchronization in distributed clusters. Our approach enables a fast and scalable inference serving system by employing an offline planner for joint computation allocation and communication scheduling, along with an online scheduler for dynamic traffic management and load balancing. We implement a prototype on a testbed comprising six servers and two programmable switches. Experimental results demonstrate that HeroServe improves scalability by$1.53 \times$while achieving lower latency compared to state-of-the-art solutions.

In this paper, we present ROCK, a novel system for efficiently serving thousands of LoRA adapters for multimodal models in cloud environments. Through extensive analysis of production workloads, we identify key challenges in current cloud-based image generation services: extreme request burstiness (up to$90 \times$normal rates), heterogeneous task characteristics, and inefficient adapter management that wastes 40 % of GPU memory and increases delays by$3 x$during peak times. ROCK addresses these challenges through a three-layer architecture that decouples hardware, adapters, and requests. Our system features dynamic heterogeneous queues that match tasks to appropriate resources based on multidimensional feature vectors, and a multilevel orchestration framework that intelligently manages adapter placement across heterogeneous storage. Experiments on a 64-GPU testbed demonstrate that ROCK reduces average response latency by$16-26 \%$, and achieves an 84.1 % cache hit rate for LoRA adapters-outperforming traditional approaches while reducing adapter update frequency by up to 77 %.

Model binarization has made significant progress in enabling real-time and energy-efficient computation for con-volutional neural networks (CNN), offering a potential solution to the deployment challenges faced by Vision Transformers (ViTs) on edge devices. However, due to the structural differences between CNN and Transformer architectures, simply applying binary CNN strategies to the ViT models will lead to a significant performance drop. To tackle this challenge, we propose BHViT, a binarization-friendly hybrid ViT architecture and its full binarization model with the guidance of three important observations. Initially, BHViT utilizes the local information interaction and hierarchical feature aggregation technique from coarse to fine levels to address redundant computations stemming from excessive tokens. Then, a novel module based on shift operations is proposed to enhance the performance of the binary Multi-Layer Perceptron (MLP) module without significantly increasing computational overhead. In addition, an innovative attention matrix binarization method based on quantization decomposition is proposed to evaluate the token’s importance in the binarized attention matrix. Finally, we propose a regularization loss to address the inadequate optimization caused by the incompatibility between the weight oscillation in the binary layers and the Adam Optimizer. Extensive experimental results demonstrate that our proposed algorithm achieves SOTA performance among binary ViT methods. The source code is released at: https://github.com/IMRL/BHViT.

In the age of pervasive machine learning applications, protecting digital content from unauthorized use has become a pressing concern. Unlearnable examples (UEs)—data modified with imperceptible perturbations to inhibit model training while preserving human usability—have emerged as a promising approach. However, existing UE methods assume unauthorized trainers have extensive exposure to UEs or that models are trained from scratch, which may not hold in practical scenarios, This paper investigates the effectiveness of UEs under the few-shot learning paradigm, pitching it against prompt learning (PL) models that leverage pretrained vision-language models (VLMs), like CLIP, capable of generalizing to new classes with minimal data. To address this, we introduce an adaptive UE framework to generate unlearnable examples that specifically target the PL process. In addition, we propose a novel UE countermeasure, A3, with cross-modal adversarial feature alignment, specifically designed to circumvent UEs under few-shot PL. Experimental evaluations on 7 datasets show that A3outperforms existing PL methods, achieving up to 33% higher performance in learning from UEs. For example, in the scenario involving ω→-bounded EM perturbations, A3has an average harmonic mean accuracy across 7 datasets of 82.43%, compared to CoCoOp’s baseline of 65.47%. Our findings highlight the limitations of existing UEs against PL and lay the foundation for future data protection mechanisms.

Autonomous driving evaluation requires simulation environments that closely replicate actual road conditions, including real-world sensory data and responsive feedback loops. However, many existing simulations need to predict waypoints along fixed routes on public datasets or synthetic photorealistic data, i.e., open-loop simulation usually lacks the ability to assess dynamic decision-making. While the recent efforts of closed-loop simulation offer feedback-driven environments, they cannot process visual sensor inputs or produce outputs that differ from real-world data. To address these challenges, we propose DrivingSphere, a realistic and closed-loop simulation framework. Its core idea is to build 4D world representation and generate real-life and controllable driving scenarios. In specific, our framework includes a Dynamic Environment Composition module that constructs a detailed 4D driving world with a format of occupancy equipping with static backgrounds and dynamic objects, and a Visual Scene Synthesis module that transforms this data into high-fidelity, multi-view video outputs, ensuring spatial and temporal consistency. By providing a dynamic and realistic simulation environment, DrivingSphere enables comprehensive testing and validation of autonomous driving algorithms, ultimately advancing the development of more reliable autonomous cars. The benchmark will be publicly released.

Transformer models face scalability challenges in causal language modeling (CLM) due to inefficient memory allocation for growing keyvalue (KV) caches, which strains compute and storage resources.Existing methods like Grouped Query Attention (GQA) and tokenlevel KV optimization improve efficiency but rely on rigid resource allocation, often discarding "low-priority" tokens or statically grouping them, failing to address the dynamic spectrum of token importance.We propose mixSGA, a novel mixture-of-expert (MoE) approach that dynamically optimizes token-wise computation and memory allocation.Unlike prior approaches, mixSGA retains all tokens while adaptively routing them to specialized experts with varying KV group sizes, balancing granularity and efficiency.Our key novelties include: (1) a token-wise expert-choice routing mechanism guided by learned importance scores, enabling proportional resource allocation without token discard; (2) weight-sharing across grouped attention projections to minimize parameter overhead; and (3) an auxiliary loss to ensure one-hot routing decisions for training-inference consistency in CLMs.Extensive evaluations across Llama3, TinyLlama, OPT, and Gemma2 model families show mixSGA's superiority over static baselines.On instruction-following and continued pretraining tasks, mixSGA achieves higher ROUGE-L and lower perplexity under the same KV budgets.

An Wang, Xingwu Sun, Ruobing Xie, Shuaipeng Li, Jiaqi Zhu, Zhen Yang, Pinxue Zhao, Weidong Han, Zhanhui Kang, Di Wang, Naoaki Okazaki, Cheng-zhong Xu. Proceedings of the 2025 Conference on Empirical Methods in Natural Language Processing. 2025.

Deep learning (DL) inference services are widely recognized as crucial workloads in large-scale cloud clusters. However, due to the stringent latency requirements, cloud providers often over-provision GPU resources, resulting in underutilization of the available GPU potential. Although co-locating tasks on the same device can enhance utilization, ensuring Service Level Objectives (SLOs) guarantees for multiplexing highly dynamic inference services becomes extremely challenging due to significant resource interference.

Microservice applications are commonly deployed alongside other services to enhance resource utilization. However, this practice also leads to notable resource contention. While existing studies primarily focus on scaling critical microservices responsible for performance degradation to mitigate violations of SLAs regarding end-to-end latency in highly interfered environments, they often overlook the potential advantages of scaling non-critical microservices for optimized resource efficiency. In this paper, we introduce Grad, an intelligent microservice scaling framework by harnessing resource fungibility between critical and non-critical microservices. Addressing the challenges posed by the dynamic nature of resource fungibility during scaling, Grad incorporates three key components. First, Grad employs a modular learning approach to profile individual microservice latency in relation to environmental conditions. Utilizing gradient extracts from this profile, Grad designs a scalable optimization module to dynamically select the optimal set of microservices for scaling. To rapidly mitigate SLA violations, Grad also deploys an accurate end-to-end latency predictor, serving as an simulator to obtain real-time feedback. We evaluate Grad in our cluster using real microservice benchmarks and production traces, demonstrating its ability to reduce resource usage by $\mathbf{4 9. 1 \%}$ and lower the probability of SLA violations by $3.7 \times$ when compared to state-of-the-art solutions.

Occlusion is a key factor leading to detection failures. This paper proposes a motion-assisted detection (MAD) method that actively plans an executable path, for the robot to observe the target at a new viewpoint with potentially reduced occlusion. In contrast to existing MAD approaches that may fail in cluttered environments, the proposed framework is robust in such scenarios, therefore termed clutter resilient occlusion avoidance (CROA). The crux to CROA is to minimize the occlusion probability under polyhedron-based collision avoidance constraints via the convex-concave procedure and duality-based bilevel optimization. The system supports lidar-based MAD with intertwined execution of learning-based detection and optimization-based planning. Experiments show that CROA outperforms various MAD schemes under a sparse convolutional neural network detector, in terms of point density, occlusion ratio, and detection error, in a multi-lane urban driving scenario.

In this paper, we propose Binarized Change Detection (BiCD), the first binary neural network (BNN) designed specifically for change detection. Conventional network binarization approaches, which directly quantize both weights and activations in change detection models, severely limit the network's ability to represent input data and distinguish between changed and unchanged regions. This results in significantly lower detection accuracy compared to real-valued networks. To overcome these challenges, BiCD enhances both the representational power and feature separability of BNNs, improving detection performance. Specifically, we introduce an auxiliary objective based on the Information Bottleneck (IB) principle, guiding the encoder to retain essential input information while promoting better feature discrimination. Since directly computing mutual information under the IB principle is intractable, we design a compact, learnable auxiliary module as an approximation target, leading to a simple yet effective optimization strategy that minimizes both reconstruction loss and standard change detection loss. Extensive experiments on street-view and remote sensing datasets demonstrate that BiCD establishes a new benchmark for BNN-based change detection, achieving state-of-the-art performance in this domain.

Low-precision training is considered an effective strategy for reducing both training and downstream inference costs. Previous scaling laws for precision mainly focus on integer quantization, which pay less attention to the constituents in floating-point (FP) quantization, and thus cannot well fit the LLM losses in this scenario. In contrast, while FP quantization training is more commonly implemented in production, it's research has been relatively superficial. In this paper, we thoroughly explore the effects of FP quantization targets, exponent bits, mantissa bits, and the calculation granularity of the scaling factor in FP quantization training performance of LLM models. In addition to an accurate FP quantization unified scaling law, we also provide valuable suggestions for the community: (1) Exponent bits contribute slightly more to the model performance than mantissa bits. We provide the optimal exponent-mantissa bit ratio for different bit numbers, which is available for future reference by hardware manufacturers; (2) We discover the formation of the critical data size in low-precision LLM training. Too much training data exceeding the critical data size will inversely bring in degradation of LLM performance; (3) The optimal FP quantization precision is directly proportional to the computational power, but within a wide computational power range. We estimate that the best cost-performance precision should lie between 4-8 bits.

Accurate trajectory prediction has long been a major challenge for autonomous driving (AD). Traditional data-driven models predominantly rely on statistical correlations, often overlooking the causal relationships that govern traffic behavior. In this paper, we introduce a novel trajectory prediction framework that leverages causal inference to enhance predictive robustness, generalization, and accuracy. By decomposing the environment into spatial and temporal components, our approach identifies and mitigates spurious correlations, uncovering genuine causal relationships. We also employ a progressive fusion strategy to integrate multimodal information, simulating human-like reasoning processes and enabling real-time inference. Evaluations on five real-world datasets—ApolloScape, nuScenes, NGSIM, HighD, and MoCAD—demonstrate our model's superiority over existing state-of-the-art (SOTA) methods, with improvements in key metrics such as RMSE and FDE. Our findings highlight the potential of causal reasoning to transform trajectory prediction, paving the way for robust AD systems.

Navigating autonomous vehicles in open scenarios is a challenge due to the difficulties in handling unseen objects. Existing solutions either rely on small models that struggle with generalization or large models that are resource-intensive. While collaboration between the two offers a promising solution, the key challenge is deciding when and how to engage the large model. To address this issue, this paper proposes opportunistic collaborative planning (OCP), which seamlessly integrates efficient local models with powerful cloud models through two key innovations. First, we propose large vision model guided model predictive control (LVM-MPC), which leverages the cloud for LVM perception and decision making. The cloud output serves as a global guidance for a local MPC, thereby forming a closed-loop perception-to-control system. Second, to determine the best timing for large model query and service, we propose collaboration timing optimization (CTO), including object detection confidence thresholding (ODCT) and cloud forward simulation (CFS), to decide when to seek cloud assistance and when to offer cloud service. Extensive experiments show that the proposed OCP outperforms existing methods in terms of both navigation time and success rate.

Federated Learning (FL) enables multiple devices to collaboratively train a shared model while preserving data privacy. Most existing research assumes that all participating devices have sufficient resources to support the training process. However, the high memory requirements of model training present a significant challenge to deploying FL on resource-constrained devices in practical scenarios. To this end, this paper presents ProFL, a new framework that effectively addresses the memory constraints in FL. Rather than updating the full model during local training, ProFL partitions the model into blocks based on its original architecture and trains each block in a progressive fashion. It first trains the front blocks and safely freezes them after convergence. Training of the next block is then triggered. This process progressively grows the model to be trained until the training of the full model is completed. In this way, the peak memory footprint is effectively reduced for feasible deployment on heterogeneous devices. In order to preserve the feature representation of each block, the training process is divided into two stages: model shrinking and model growing. During the model shrinking stage, we meticulously design corresponding output modules to assist each block in learning the expected feature representation and obtain the initialization model parameters. Subsequently, the obtained output modules and initialization model parameters are utilized in the corresponding model growing stage, which progressively trains the full model. Additionally, a novel metric from the scalar perspective is proposed to assess the learning status of each block, enabling us to securely freeze it after convergence and initiate the training of the next one. Finally, we theoretically prove the convergence of ProFL and conduct extensive experiments on representative models and datasets to evaluate its effectiveness. The results demonstrate that ProFL effectively reduces the peak memory footprint by up to 57.4% and improves model accuracy by up to 82.4%.

With the rapid development of large language models (LLMs), there has been a strong interest in integrating other modalities such as image comprehension capabilities. While they have shown impressive performance in various multimodal tasks, the robustness of Visual Language Models (VLMs) has not been thoroughly investigated. We mainly focus on the robustness of VLMs on visual adversarial examples. In this work, we explore the capability of adversarial examples targeting VLMs. We highlight that the multimodal nature of VLMs presents a unique attack surface to manipulate the outputs of the LLMs, and the continuous nature of visual inputs further enhances the effectiveness of adversarial attacks against language generative models. Furthermore, we demonstrate three application scenarios for adversarial examples targeting VLMs: image description, jailbreaking, and information hiding. We conduct experiments on several leading open-source VLMs and demonstrate the successful application of adversarial examples in all the proposed scenarios. We hope that our findings would enable the development of multimodal models more robust to adversarial attacks. Our code is available at https://github.com/lafeat/m3-break.

Industrial anomaly detection for 2D objects has gained significant attention and achieved progress in anomaly detection (AD) methods. However, identifying 3D depth anomalies using only 2D information is insufficient. Despite explicitly fusing depth information into RGB images or using point cloud backbone networks to extract depth features, both approaches struggle to adequately represent 3D information in multimodal scenarios due to the disparities among different modal information. Additionally, due to the scarcity of abnormal samples in industrial data, especially in multimodal scenarios, it is necessary to perform anomaly generation to simulate real-world abnormal samples. Therefore, we propose a novel unified multimodal anomaly detection framework to address these issues. Our contributions consist of 3 key aspects. (1) We extract visible depth information from 3D point cloud data simply and use 2D RGB images to represent appearance, which disentangles depth and appearance to support unified anomaly generation. (2) Benefiting from the flexible input representation, the proposed Multi-Scale Gaussian Anomaly Generator and Unified Texture Anomaly Generator can generate richer anomalies in RGB and depth. (3) All modules share parameters for both RGB and depth data, effectively bridging 2D and 3D anomaly detection. Subsequent modules can directly leverage features from both modalities without complex fusion. Experiments show our method outperforms state-of-the-art (SOTA) on MVTec-3D AD and Eyecandies datasets. Code available at: https://github.com/Xantastic/BridgeNet

The increasing emphasis on privacy and data security has driven the adoption of federated learning (FL). Prompt learning (PL), which fine-tunes prompt embeddings of pretrained models, has gained a surge of interest in FL community, marked by the emergence of an influx of federated prompt learning (FPL) algorithms. Despite recent advancements, a systematic understanding of their underlying mechanisms and principled guidelines for deploying these techniques in different FL scenarios remain absent. Moreover, inconsistent experimental protocols, limited evaluation scenarios, and the lack of the proper assessment of centralized PL methods in existing works have obscured the essence of these algorithms. To close these gaps, we introduce a comprehensive benchmark, named F LIP, to achieve standardized FPL evaluation. F LIP assesses the performance of 13 centralized and FPL methods across 3 FL protocols and 12 open datasets, considering 6 distinct evaluation scenarios. Our findings demonstrate that PL maintains strong generalization performance in both in-distribution and out-of-distribution settings with minimal resource consumption, but there is no silver bullet found for diverse FPL scenarios. The results (1) pinpoint the suitable application scenarios of each FPL algorithm, (2) demonstrate the competitiveness of adapted centralized PL methods, and (3) offer notable insights to interpret their effectiveness and remaining challenges. All benchmarks and code are available to facilitate further research in this domain.

Synthetic data is crucial for advancing autonomous driving (AD) systems, yet current state-of-the-art video generation models, despite their visual realism, suffer from subtle geometric distortions that limit their utility for downstream perception tasks. We identify and quantify this critical issue, demonstrating a significant performance gap in 3D object detection when using synthetic versus real data. To address this, we introduce Reinforcement Learning with Geometric Feedback (RLGF), RLGF uniquely refines video diffusion models by incorporating rewards from specialized latent-space AD perception models. Its core components include an efficient Latent-Space Windowing Optimization technique for targeted feedback during diffusion, and a Hierarchical Geometric Reward (HGR) system providing multi-level rewards for point-line-plane alignment, and scene occupancy coherence. To quantify these distortions, we propose GeoScores. Applied to models like DiVE on nuScenes, RLGF substantially reduces geometric errors (e.g., VP error by 21\%, Depth error by 57\%) and dramatically improves 3D object detection mAP by 12.7\%, narrowing the gap to real-data performance. RLGF offers a plug-and-play solution for generating geometrically sound and reliable synthetic videos for AD development.

Deploying large language models (LLMs) on edge devices enables real-time responses while preserving user privacy. However, constrained memory and compute resources pose significant challenges for high-quality, single-device inference. To address this, we propose HCInfer, a hierarchical coordination framework for collaborative LLM inference across edge devices. By leveraging idle neighboring devices, HCInfer alleviates performance bottlenecks typical in isolated deployments. HCInfer employs a two-level coordination strategy. At the inter-device level, it leverages idle neighboring devices to collaboratively process attention computations, significantly reducing synchronization overhead. At the intra-device level, it applies finegrained memory and compute optimizations to fully exploit local hardware capabilities. Building on this architecture, HCInfer integrates three key components: (1) Asymmetric Transformer decomposition decouples attention and FFN computation, enabling selective and parallel execution across devices. (2) Layer-wise subdeadline scheduling dynamically profiles execution latency and adapts precision or structure to meet real-time constraints (3) An Overhead Mitigation Module efficiently manages on-device resource usage to support scalability without overwhelming hardware. We evaluate HCInfer on PC, smart home, and mobile platforms using OPT-13B, Qwen2.5-14B, and Llama2-13B models. Experiments show HCInfer achieves 1.67× to 4.3× speedup in TTFT and 1.16× to 17.15× speedup in TPOT compared to existing baselines, maintaining a sub-deadline miss rate (SubDMR) of 15.3% under worst-case conditions while keeping model accuracy degradation within 8% for typical cases and up to 11% in extreme scenarios. These results demonstrate HCInfer's potential to enable efficient and responsive LLM inference in real-world edge environments.

The significant resource demands in LLM serving prompts production clusters to fully utilize heterogeneous hardware by partitioning LLM models across a mix of high-end and low-end GPUs. However, existing parallelization approaches often struggle to scale efficiently in heterogeneous environments due to their coarse-grained and static parallelization strategies.

Human activity recognition (HAR) utilizing WiFi channel state information (CSI) holds profound implications owing to the pervasive WiFi coverage in daily life. Deep learning has enabled the development of many high-precision HAR systems, but also brings the challenges of degrading performance in the new environments and high complexity issues for implementing on a single chip. In this paper, we introduce a lightweight HAR scheme for rapid adaption to the new environments, which can be implemented in a WiFi based artificial intelligence internet of things (AIoT) chip. The proposed scheme consists of two modules, which are change pattern extraction (CPE) and self-attention based adaptive model (SAAM). In CPE, the intricate multi-subcarrier CSI are transformed into unique change patterns closely related to each human activity with low complexity using deep non-negative matrix factorization (DNMF). Then, SAAM facilitates the correlation of change patterns across different environments through learned temporal features, enabling rapid generalization to new environments with only a few new samples, boasting advantages of low training costs, parameter memory usage, and computational time. Experimental results demonstrate that our system achieves not only 94% accuracy in the original environment, but also exhibits promising performance in new environments, requiring only three new training samples for each activity.

In 6G vehicle to everything (V2X) communications, accurate localization is the foundation of high quality intelligent services to users. However, accurate localization requires a significant amount of information and should overcome the Nonline of Sight (NLoS). In this article, we propose the reconfigurable-intelligent-surface (RIS)-assisted cooperative channel state information localization (RCSIL) system to analyze and optimize the localization accuracy based on effective channel state information (CSI). RIS improves the channel quality by adjusting the phase of the reflection unit, thereby affecting the localization performance. Using RIS as an anchor saves a lot of hardware resources. However, the optimization problem of the RIS phase modulation scheme is a nonconvex nonlinear integer programming which is complex and hard to solve. Thus, we propose an Adagrad-gradient-descent-based phase optimization (AG-PO) algorithm, which computes the optimal phase of each reflecting unit in parallel. AG-PO has low complexity, effectively improves positioning accuracy, and performs fast computing. The simulation results indicate that AG-PO outperforms alternating optimization (AO) and genetic algorithm (GA) with a 44.17% and 44.02% reduction in squared position error bound (SPEB) respectively. Thus, RCSIL can be widely applied in 6G V2X communications systems.

Due to the rich information and original data distribution, RAW data are widely used in many computer vision applications. However, the use of RAW video remains limited because of the high storage costs associated with data collection. Previous works have attempted to reconstruct RAW frames from sRGB data using small sampled metadata from the original RAW frames. Yet, these algorithms struggle with RAW video reconstruction due to the high computational cost of sampling metadata on cameras. To address these issues, we propose a new RAW video reconstruction pipeline that de-renders high-quality RAW videos from sRGB data using only one initial RAW frame as a reference. Specifically, we introduce three new models to achieve this goal. First, we present the Temporal-Affinity Guided De-rendering Network. This network leverages the temporal affinity between adjacent frames to construct a reference RAW image from previous RAW pixels. The corresponding RAW pixels in the previous frame provide valuable information about the original RAW data distribution, aiding in the precise reconstruction of the current frame. Second, to recover the missing RAW pixels caused by camera and foreground movement, we fully exploit the rich prior information from a pre-trained diffusion model and propose the RAW In-painting Model. This model can accurately fill in hollow regions in a RAW image based on the corresponding sRGB image and the surrounding RAW context. Lastly, we present a lightweight content-aware video clipper that automatically adjusts the clip length used for RAW video reconstruction, thereby balancing storage requirements with reconstruction quality. To better evaluate the performance of the proposed framework across different devices, we introduce the first RAW video reconstruction benchmark that comprises RAW videos from six types of camera devices with challenging scenarios. Experimental results demonstrate that our algorithm can accurately reconstruct RAW videos across all the scenarios.

Addressing the pervasive challenge of imperfect data in autonomous vehicle (AV) systems, this study pioneers an integrated trajectory prediction model, WAKE, that fuses physics-informed methodologies with sophisticated machine learning techniques. Our model operates in two principal stages: the initial stage utilizes a Wavelet Reconstruction Network to accurately reconstruct missing observations, thereby preparing a robust dataset for further processing. This is followed by the Kinematic Bicycle Model which ensures that reconstructed trajectory predictions adhere strictly to physical laws governing vehicular motion. The integration of these physics-based insights with a subsequent machine learning stage, featuring a Quantum Mechanics-Inspired Interaction-aware Module, allows for sophisticated modeling of complex vehicle interactions. This fusion approach not only enhances the prediction accuracy but also enriches the model's ability to handle real-world variability and unpredictability. Extensive tests using specific versions of MoCAD, NGSIM, HighD, INTERACTION, and nuScenes datasets featuring missing observational data, have demonstrated the superior performance of our model in terms of both accuracy and physical feasibility, particularly in scenarios with significant data loss-up to 75% missing observations. Our findings underscore the potency of combining physics-informed models with advanced machine learning frameworks to advance autonomous driving technologies, aligning with the interdisciplinary nature of information fusion.

ABSTRACT Objective The microservices architecture has become a dominant paradigm in cloud computing due to its advantages in development, deployment, modularity, and scalability. Ensuring Quality of Service (QoS) through efficient Service Level Objective (SLO) resource allocation is a critical challenge. Current frameworks for microservice autoscaling based on SLOs often rely on heavy and complex models that are time‐consuming and resource‐intensive, making them unsuitable for rapidly changing environments and highly dynamic workloads. Methods This study proposes LSRAM (Lightweight SLO Resource Allocation Management), a novel framework designed to overcome the limitations of existing SLO‐based autoscaling methods. LSRAM operates in two stages: 1). Lightweight SLO Resource Allocation Model: Computes optimal SLO resource allocation for each microservice using a gradient descent method, ensuring rapid computation and minimal computational overhead. 2). SLO Resource Update Model: Adapts resource allocation dynamically in response to changes in the cluster environment, such as varying loads and application types, without requiring extensive retraining. Results LSRAM effectively addresses scenarios involving bursty traffic and fluctuating workloads. Compared to state‐of‐the‐art SLO allocation frameworks, LSRAM achieves the following: 1). Reduces resource usage by 17%. 2). Maintains QoS guarantees for users, even under dynamic conditions. 3). Demonstrates faster adaptability to changes in the system environment due to its lightweight design. Conclusion LSRAM offers a scalable, efficient, and adaptive solution for SLO‐based resource allocation in microservices architectures. By reducing resource usage while maintaining QoS, it provides a robust framework for managing dynamic and unpredictable workloads in cloud environments. Its lightweight design ensures practical applicability and superior performance compared to traditional, resource‐intensive methods.

ABSTRACT Background Cloud‐native applications are increasingly becoming popular in modern software design. Employing a microservice‐based architecture into these applications is a prevalent strategy that enhances system availability and flexibility. However, cloud‐native applications introduce new challenges, including frequent inter‐service communication and the management of heterogeneous codebases and hardware, resulting in unpredictable complexity and dynamism. Furthermore, as applications scale, only limited research teams or enterprises possess the resources for large‐scale deployment and testing, which impedes progress in the cloud‐native domain. Aims To address these challenges, we propose CloudNativeSim, a simulator for cloud‐native applications with a microservice‐based architecture. Results CloudNativeSim offers several key benefits: (i) comprehensive and dynamic modeling for cloud‐native applications, (ii) an extended simulation framework with new policy interfaces for scheduling cloud‐native applications, and (iii) support for customized application scenarios and user feedback based on Quality of Service (QoS) metrics. Conclusion CloudNativeSim can be easily deployed on standard computers to manage a high volume of requests and services. Its performance was validated through a case study, demonstrating higher than 94.5% accuracy in terms of response time simulation. The study further highlights the feasibility of CloudNativeSim by illustrating the effects of various scaling policies.

Microservice architecture has transformed traditional monolithic applications into lightweight components. Scaling these lightweight microservices is more efficient than scaling servers. However, scaling microservices still faces the challenges resulting from the unexpected spikes or bursts of requests, which are difficult to detect and can degrade performance instantaneously. To address this challenge and ensure the performance of microservice-based applications, we propose a status-aware and elastic scaling framework called StatuScale , which is based on load status detector that can select appropriate elastic scaling strategies for differentiated resource scheduling in vertical scaling. Additionally, StatuScale employs a horizontal scaling controller that utilizes comprehensive evaluation and resource reduction to manage the number of replicas for each microservice. We also present a novel metric named correlation factor to evaluate the resource usage efficiency. Finally, we use Kubernetes, an open source container orchestration and management platform, and realistic traces from Alibaba to validate our approach. The experimental results have demonstrated that the proposed framework can reduce the average response time in the Sock-Shop application by 8.59% to 12.34% and in the Hotel-Reservation application by 7.30% to 11.97%, decrease service level objective violations, and offer better performance in resource usage compared to baselines.

Distributed machine learning over geo-distributed clouds enables joint training of data located in different regions, alleviating the burden of transferring large volumes of training datasets, which greatly saves bandwidth. However, the limited capacity of WAN links slows down the inter-cloud communications, which significantly decelerates the synchronization of distributed machine learning over geo-distributed clouds. Besides, the multi-tenancy in clouds results in multiple training tasks running simultaneously, whose synchronizations consistently compete for the limited WAN bandwidth with each other, which further aggravates the training performance of each task. While existing works optimize synchronizations through techniques like gradient compression, multi-resource interleaving and so on, none of them targets at the synchronization congestion especially due to multi-tenant learning, which results in inferior training performance.To solve these problems, we propose a simple but effective scheme, SCC, for fast and efficient multi-tenant learning via synchronization congestion control. SCC monitors the cross-cloud network conditions and evaluates the synchronization congestion level based on the round-trip transmission time for each synchronization. Then SCC alleviates synchronization congestion via controlling the synchronization frequency according to the synchronization congestion level in a probabilistic way. Extensive experiments are conducted within our testbeds consisted of 16 NVIDIA V100 GPUs to evaluate the performance of SCC, and comparison results show that SCC can reduce the average training completion time and makespan by up to 28.6% and 43.2% over SAP-SGD [1]. Targeted experiments are conducted to demonstrate the effectiveness and robustness of SCC.

Efficient resource utilization is crucial in real-world applications, especially for balancing loads across machines handling specific job types. This paper introduces a novel batch-ordered job-store scheduling model, where jobs in a batch are scheduled sequentially, with their operations allocated in a round-robin fashion across two scenarios. We establish that this problem is NP-hard and analyze it in both online and offline settings. In the online case, we first examine the exclusive scenario, where operations within the same job must be scheduled on different machines, and show that a load greedy (LG) algorithm achieves a tight competitive ratio of$2 - \frac{1}{m}$, withmrepresenting the number of machines. Next, we consider the circular scenario, which requires maintaining the circular order of operations across ordered machines. In this context, we analyze potential anomalies in load distribution during local optimality achieved by the ordered load greedy (OLG) algorithm and provide bounds on the occurrence of these anomalies and the maximum load in each local scheduling round. In the offline case, we abstract each OLG scheduling process as a generalized circular sequence alignment (CSA) problem and develop a dynamic programming-based matching (DPM) algorithm to solve it. To further enhance load balancing, we develop a dynamic programming-based optimization (DPO) algorithm to schedule multiple jobs simultaneously in both scenarios. Experimental results confirm the efficiency of DPM for the CSA problem, and we validate the load balancing effectiveness of both online and offline algorithms using real traffic datasets. These theoretical findings and algorithmic implementations lay a solid groundwork for future practical advancements.

This article is concerned with the secure collision-free tracking control problem for autonomous vehicles with uncertainties, where system signals are transmitted through constrained communication networks. In such open and uncertain environments, the control performance of vehicles is seriously affected by privacy leakage, non-Gaussian noise, and obstacles. The aim of this research is to propose a tracking control scheme that ensures security, mean-square boundedness, and collision-free performance concurrently. Initially, to safeguard the privacy of the transmitted data and to enable secure tracking control, a dynamic encoding-based ElGamal encryption mechanism is introduced, which is further embedded in the design of the observer-based tracking controller. Subsequently, a collision-free chance-constrained index is proposed for achieving real-time obstacle avoidance by comprehensively considering the influence of stochastic noises. A thorough analysis is conducted to examine the impact of non-Gaussian noise and unmeasurable states on the performance of collision-free tracking control. Sufficient conditions are derived to guarantee the desired performance, and the corresponding control inputs are obtained by solving certain optimization problems subject to chance constraints. Finally, an illustrative example is provided to validate the effectiveness of the proposed secure collision-free tracking controller.

Most autonomous driving (AD) datasets incur substantial costs for collection and labeling, inevitably yielding a plethora of low-quality and redundant data instances, thereby compromising performance and efficiency. Many applications in AD systems necessitate high-quality training datasets using both existing datasets and newly collected data. In this paper, we propose a traffic scene joint active learning (TSceneJAL) framework that can efficiently sample the balanced, diverse, and complex traffic scenes from both labeled and unlabeled data. The novelty of this framework is threefold: 1) a scene sampling scheme based on a category entropy, to identify scenes containing multiple object classes, thus mitigating class imbalance for the active learner; 2) a similarity sampling scheme, estimated through the directed graph representation and a marginalize kernel algorithm, to pick sparse and diverse scenes; 3) an uncertainty sampling scheme, predicted by a mixture density network, to select instances with the most unclear or complex regression outcomes for the learner. Finally, the integration of these three schemes in a joint selection strategy yields an optimal and valuable subdataset. Experiments on the KITTI, Lyft, nuScenes and SUScape datasets demonstrate that our approach outperforms existing state-of-the-art methods on 3D object detection tasks with up to 12% improvements.

Online map matching (MM) aligns real-time GPS trajectories with digital road networks, playing a vital role in vehicle navigation, route planning, and traffic analysis. Hidden Markov Models (HMMs) are widely used for their interpretability and ability to handle low GPS sampling rates. However, in urban scenarios characterized by complex road networks, significant GPS localization error, and dynamic traffic conditions, existing HMM-based methods face challenges such as large road search spaces due to uniform GPS localization error distributions (GLED) and inaccurate route accessibility estimates stemming from inadequate consideration of real-time traffic conditions. This paper proposes an improved HMM-based MM method, recognizing that urban vehicle trajectories can be categorized into two types: path-free (e.g., taxis, private cars) and path-constrained (e.g., buses). Analyzing path-constrained trajectories helps estimate fine-grained GLED and real-time traffic states of path-free vehicles more precisely. The novelty of our approach lies in two aspects: i) Using a hierarchical spectral clustering algorithm based on GPS localization errors of path-constrained bus trajectories, a city is divided into fine-grained sub-regions with consistent GLED. This enables the HMM an adaptive road search scopes, improving online MM efficiency. ii) Gradient boosting trees, known for their interpretability, estimate free-flow speeds by integrating path-constrained trajectories with the factors like road attributes and time, optimizing HMM state transition probabilities for path-free trajectory MM. Experiments on real-world data demonstrate that the optimized HMM methods, leveraging different trajectory types, significantly enhance MM efficiency and accuracy compared to baseline models.The codebase of our methods and datasets are available at https://github.com/jacklee018/onlineMM-IPCT.

Accurately anticipating traffic accidents is a fundamental task for the safe and effective deployment of autonomous vehicles (AVs). However, existing models primarily rely on dashcam footage and often fail to generalize across varied driving scenarios due to their dependence on visual data and the rarity of high-risk events in datasets. These limitations undermine their robustness and reduce practical applicability in dynamic, unpredictable environments. To address these challenges, this study proposes a novel approach, termed MLTA, which integrates multimodal learning with the hypergraph attention network to hierarchically extract and capture cross-modal interaction. It leverages LLava-next, a multimodal large language model (MLLM) guided by the Chain-of-Thought (CoT) prompting paradigm, to produce context-aware interpretations of traffic scenes. This is further enhanced by a human-inspired attention mechanism that mimics the decision-making priorities of experienced human drivers. This combination enables more accurate identification of critical elements in a scene, improving both prediction precision and timeliness. Extensive experiments on four real-world datasets—DAD, A3D, CCD, and DADA-2000—show that our approach consistently outperforms state-of-the-art (SOTA) methods, demonstrating strong adaptability and robustness in complex driving environments.

In mixed autonomous driving environments, accurately predicting the future trajectories of surrounding vehicles is crucial for the safe operation of autonomous vehicles (AVs). In driving scenarios, a vehicle’s trajectory is determined by the decision-making process of human drivers. However, existing models primarily focus on the inherent statistical patterns in the data, often neglecting the critical aspect of understanding the decision-making processes of human drivers. This oversight results in models that fail to capture the true intentions of human drivers, leading to suboptimal performance in long-term trajectory prediction. To address this limitation, we introduce a Cognitive-Informed Transformer (CITF) that incorporates a cognitive concept, Perceived Safety, to interpret drivers’ decision-making mechanisms. Perceived Safety encapsulates the varying risk tolerances across drivers with different driving behaviors. Specifically, we develop a Perceived Safety-aware Module that includes a Quantitative Safety Assessment for measuring the subject risk levels within scenarios, and Driver Behavior Profiling for characterizing driver behaviors. Furthermore, we present a novel module, Leanformer, designed to capture social interactions among vehicles. CITF demonstrates significant performance improvements on three well-established datasets. In terms of long-term prediction, it surpasses existing benchmarks by 12.0% on the NGSIM, 28.2% on the HighD, and 20.8% on the MoCAD dataset. Additionally, its robustness in scenarios with limited or missing data is evident, surpassing most state-of-the-art (SOTA) baselines, and paving the way for real-world applications.

As an emerging technology, Digital Twin (DT) can provide a virtual representation of transportation infrastructures to achieve efficient and precise management of Intelligent Transportation Systems (ITS). However, a mixed traffic scenario of coexisting intelligent connected vehicles (ICVs) and non-intelligent connected vehicles (N-ICVs) increases challenges for digital ITS. N-ICVs are unable to generate and update their DT models independently due to constrained communication and computing capabilities. It is crucial to achieve real-time DT model update and migration of N-ICVs. In this paper, we propose a cooperative perception aided DT model update and migration approach, which dispatches ICVs to cooperatively sense and transmit information of nearby N-ICVs to assist in generating N-ICVs’ DT models. In particular, with the objective of minimizing the average maximum weighted age of information (AMWAoI), we jointly optimize the cooperative ICV selection as well as the bandwidth and computation allocations while guaranteeing the perception performance. We then propose a sensing data weighted size maximization matching algorithm to achieve an optimal ICV selection strategy, and the bandwidth and computation allocations are optimized by the gradient descent algorithm. Considering the dynamic nature of vehicular networks, a deep reinforcement learning-based access selection and DT model migration algorithm is further proposed to achieve continuous service provisioning. Simulation results demonstrate that the proposed algorithm achieves the lowest AMWAoI while meeting the perception performance requirement.

Transport Mode Split (TMS) represents the distribution of trips among transport modes between city regions. Accurate TMS prediction is crucial for urban planning and traffic management. Traditional methods, such as curve models and discrete choice models, often fail to capture user travel mode pReferences due to dataset limitations and inadequate spatiotemporal modeling. While deep learning models have been applied in related domains such as ride-sharing and traffic prediction, they typically focus on a single transport mode and do not model the competitive relationships between modes. We propose a novel Deep Learning-based Framework for Transport Mode Split Prediction (DMSP), integrating restricted-route (bus/subway) and free-route (taxi) data. It introduces two key innovations: 1) efficient extraction of TMS samples from sparse raw data, enriched with spatiotemporal context, and 2) a prediction model extending discrete choice theory via maximum utility principles. Mode utilities are estimated using CNNs for travel mode-specific effects and GNN for spatiotemporal dependencies. Experiments on six months of Shenzhen data show that DMSP reduces MAPE by over 4% compared to existing methods.

Mobile edge computing (MEC) has revolutionized the way computational tasks are offloaded and latency is reduced by leveraging edge servers close to end devices. Efficient resource allocation and task offloading are crucial for enhancing system performance in MEC environments. Traditional reinforcement learning (RL) approaches have shown promise in optimizing resource allocation and task offloading problems. However, they often face challenges such as high computational complexity and the need for extensive training data. Quantum reinforcement learning (QRL) emerges as a promising solution to overcome these limitations by leveraging quantum computing principles to enhance efficiency and scalability. In this paper, we propose a hybrid quantum-classical non-sequential model for joint resource allocation and task offloading in MEC systems. Our model combines the advantages of RL in handling environmental dynamics and quantum computing in reducing adjustable parameters and accelerating the training process. Extensive experiments demonstrate that our proposed algorithm can achieve higher training and inference performance under various parameter settings compared to traditional RL models and previous QRL models.

Federated learning (FL) is a new learning paradigm that enables multiple clients to collaboratively train a high-performance model while preserving user privacy. However, the effectiveness of FL heavily relies on the availability of accurately labeled data, which can be challenging to obtain in real-world scenarios. To address this issue and robustly train shared models using distributed noisy labeled data, we propose FedDQ, a noise-robust FL framework that utilizes co-distillation and quality-aware aggregation techniques. FedDQ incorporates two key features: a noise-adaptive training strategy and an efficient label-correcting mechanism. The noise-adaptive training strategy relies on the estimation of labels' noise levels to dynamically adjust clients' training engagement, which mitigates the impact of wrong labels while efficiently exploring features from clean data. In addition, FedDQ designs a two-head network and employs it for co-distillation. The co-distillation strategy facilitates knowledge transfer among clients to share the representational capabilities. Besides, FedDQ enhances label correction to rectify improper labels through co-filtering and label correction. The experimental results demonstrate the effectiveness of FedDQ in improving model performance and handling noisy data challenges in FL settings. On the CIFAR-100 dataset with noisy labels, FedDQ exhibits a notable improvement of up to 32.4% compared to the baseline method.

The advent of large-scale pretrained language models (PLMs) has contributed greatly to the progress in natural language processing (NLP). Despite its recent success and wide adoption, fine-tuning a PLM often suffers from overfitting, which leads to poor generalizability due to the extremely high complexity of the model and the limited training samples from downstream tasks. To address this problem, we propose a novel and effective fine-tuning framework, named layerwise noise stability regularization (LNSR). Specifically, our method perturbs the input of neural networks with the standard Gaussian or in-manifold noise in the representation space and regularizes each layer's output of the language model. We provide theoretical and experimental analyses to prove the effectiveness of our method. The empirical results show that our proposed method outperforms several state-of-the-art algorithms, such as norm and start point (L2-SP), Mixout, FreeLB, and smoothness inducing adversarial regularization and Bregman proximal point optimization (SMART). In addition to evaluating the proposed method on relatively simple text classification tasks, similar to the prior works, we further evaluate the effectiveness of our method on more challenging question-answering (QA) tasks. These tasks present a higher level of difficulty, and they provide a larger amount of training examples for tuning a well-generalized model. Furthermore, the empirical results indicate that our proposed method can improve the ability of language models to domain generalization.

Federated learning (FL) collaboratively trains a shared global model depending on multiple local clients, while keeping the training data decentralized to preserve data privacy. However, standard FL methods ignore the noisy client issue, which may harm the overall performance of the shared model. We first investigate the critical issue caused by noisy clients in FL and quantify the negative impact of the noisy clients in terms of the representations learned by different layers. We have the following two key observations: 1) the noisy clients can severely impact the convergence and performance of the global model in FL and 2) the noisy clients can induce greater bias in the deeper layers than the former layers of the global model. Based on the above observations, we propose federated noisy client learning (Fed-NCL), a framework that conducts robust FL with noisy clients. Specifically, Fed-NCL first identifies the noisy clients through well estimating the data quality and model divergence. Then robust layerwise aggregation is proposed to adaptively aggregate the local models of each client to deal with the data heterogeneity caused by the noisy clients. We further perform label correction on the noisy clients to improve the generalization of the global model. Experimental results on various datasets demonstrate that our algorithm boosts the performances of different state-of-the-art systems with noisy clients. Our code is available at https://github.com/TKH666/Fed-NCL.

This article is concerned with the secure state estimation problem for artificial neural networks (ANNs) subject to unknown-but-bounded noises, where sensors and the remote estimator are connected via open and bandwidth-limited communication networks. Using the encoding-decoding mechanism (EDM) and the Paillier encryption technique, a novel homomorphic encryption scheme (HES) is introduced, which aims to ensure the secure transmission of measurement information within communication networks that are constrained by bandwidth. Under this encoding-decoding-based HES, the data being transmitted can be encrypted into ciphertexts comprising finite bits. The emphasis of this research is placed on the development of a secure set-membership state estimation algorithm, which allows for the computation of estimates using encrypted data without the need for decryption, thereby ensuring data security throughout the entire estimation process. Taking into account the unknown-but-bounded noises, the underlying ANN, and the adopted HES, sufficient conditions are determined for the existence of the desired ellipsoidal set. The related secure state estimator gains are then derived by addressing optimization problems using the Lagrange multiplier method. Lastly, an example is presented to verify the effectiveness of the proposed secure state estimation approach.

Accurate and robust state estimation at nighttime is essential for autonomous robotic navigation to achieve nocturnal or round-the-clock tasks. An intuitive question arises: can low-cost standard cameras be exploited for nocturnal state estimation? Regrettably, most existing visual methods may fail under adverse illumination conditions, even with active lighting or image enhancement. A pivotal insight, however, is that streetlights in most urban scenarios act as stable and salient prior visual cues at night, reminiscent of stars in deep space aiding spacecraft voyage in interstellar navigation. Inspired by this, we propose Night-Voyager, an object-level nocturnal vision-aided state estimation framework that leverages prior object maps and keypoints for versatile localization. We also find that the primary limitation of conventional visual methods under poor lighting conditions stems from the reliance on pixel-level metrics. In contrast, metric-agnostic, nonpixel-level object detection serves as a bridge between pixel-level and object-level spaces, enabling effective propagation and utilization of object map information within the system. Night-Voyager begins with a fast initialization to solve the global localization problem. By employing an effective two-stage cross-modal data association, the system delivers globally consistent state updates using map-based observations. To address the challenge of significant uncertainties in visual observations at night, a novel matrix Lie group formulation and a feature-decoupled multistate invariant filter are introduced, ensuring consistent and efficient estimation. Through comprehensive experiments in both simulation and diverse real-world scenarios (spanning approximately 12.3 km), Night-Voyager showcases its efficacy, robustness, and efficiency, filling a critical gap in nocturnal vision-aided state estimation.

Navigating a nonholonomic robot in a cluttered, unknown environment requires accurate perception and precise motion control for real-time collision avoidance. This article presents neural proximal alternating-minimization network (NeuPAN): a real-time, highly accurate, map-free, easy-to-deploy, and environment-invariant robot motion planner. Leveraging a tightly coupled perception-to-control framework, NeuPAN has two key innovations compared to existing approaches: first, it directly maps raw point cloud data to a latent distance feature space for collision-free motion generation, avoiding error propagation from the perception to control pipeline; second, it is interpretable from an end-to-end model-based learning perspective. The crux of NeuPAN is solving an end-to-end mathematical model with numerous point-level constraints using a plug-and-play proximal alternating-minimization network, incorporating neurons in the loop. This allows NeuPAN to generate real-time, physically interpretable motions. It seamlessly integrates data and knowledge engines, and its network parameters can be fine-tuned via backpropagation. We evaluate NeuPAN on a ground mobile robot, a wheel-legged robot, and an autonomous vehicle, in extensive simulated and real-world environments. Results demonstrate that NeuPAN outperforms existing baselines in terms of accuracy, efficiency, robustness, and generalization capabilities across various environments, including the cluttered sandbox, office, corridor, and parking lot. We show that NeuPAN works well in unknown and unstructured environments with arbitrarily shaped objects, transforming impassable paths into passable ones.

Trajectory prediction and planning remain key challenges for autonomous vehicles (AVs), particularly in complex and dynamic environments. Existing methods, typically based on static safety metrics like Time-to-Collision (TTC), fail to account for the evolving nature of risk in real-world traffic. This paper proposes a novel Safety-Aware Trajectory Prediction and Planning (SA-TP$^{2}$) model, which introduces an adaptive driver risk field to simulate human- like risk perception and decision-making. By dynamically modeling risk as a continuous variable, SA-TP$^{2}$adjusts vehicle trajectories in real-time, accounting for interactions with other agents, road conditions, and environmental uncertainties. The model integrates imitation learning (IL), rule-based strategies, and physics-informed neural networks (PINNs) to ensure safe, efficient, and human-compatible behavior. A Linformer-based architecture and Temporal Hypergraph Convolution Network (THGCN) are introduced to optimize computational efficiency, enabling real-time operation in resource-constrained environments. Experimental results on benchmark datasets including NGSIM, HighD, MoCAD, and NuScenes demonstrate that SA-TP$^{2}$achieves SOTA performance in trajectory prediction. Additionally, extensive closed-loop testing on the NuPlan and CommonRoad platforms further confirms that SA-TP$^{2}$outperforms existing baselines, paving the way for safer navigation of autonomous driving systems.

As Large Language Models (LLMs) continue to advance, their parameter sizes are growing exponentially-far outpacing hardware capabilities. This widening gap necessitates distributed computing through pipeline parallelism for efficient inference. However, the uneven distribution of requests across pipeline stages creates significant performance bottlenecks in real-world deployments. To address this challenge, we presentPlanck, a performance optimization framework specifically designed for distributed LLM inference.Planck implements fine-grained control through two key mechanisms: a progressive SLO allocation strategy that dynamically adjusts time constraints based on workload patterns, and stage-specific performance controllers that prevent bottlenecks before they cascade through the system. By intelligently balancing resources across pipeline stages,Planck effectively eliminates queue buildup-essentially preventing traffic congestion before it forms. Evaluation using diverse workloads in real cloud environments demonstrates thatPlanck reduces P99 tail latency by up to 18% and decreases the longest queue lengths by as much as 47.8% across pipeline stages, significantly improving both system responsiveness and resource utilization.

High-Performance Data Analytics (HPDA) combines high-performance computing (HPC) with data analytics to uncover patterns and insights in dual-intensive applications that are both data-intensive and compute-intensive. Traditional big data frameworks and HPC technologies often struggle to address these demands independently, prompting researchers to explore their integration. Spark, known for its efficient in-memory computing with RDDs, and MPI, a foundational standard in HPC, are prominent candidates for such integration. This survey explores the integration of Spark and MPI for HPDA, highlighting their potential for unified data processing and computation. We first classify application workloads and review the characteristics and limitations of traditional frameworks. Then, we analyze the challenges and requirements of integrated architectures, focusing on the specific implementations of typical middleware-level architectures. Through comparative analysis, we highlight their advantages and limitations. Finally, we present application examples, outline key challenges and future research directions, and briefly discuss progress in integration approaches for other technology combinations.

Despite the great success of large language models (LLMs) in various tasks, they suffer from generating hallucinations. We introduce Truth Forest, a method that enhances truthfulness in LLMs by uncovering hidden truth representations using multi-dimensional orthogonal probes. Specifically, it creates multiple orthogonal bases for modeling truth by incorporating orthogonal constraints into the probes. Moreover, we introduce Random Peek, a systematic technique considering an extended range of positions within the sequence, reducing the gap between discerning and generating truth features in LLMs. By employing this approach, we improved the truthfulness of Llama-2-7B from 40.8% to 74.5% on TruthfulQA. Likewise, significant improvements are observed in fine-tuned models. We conducted a thorough analysis of truth features using probes. Our visualization results show that orthogonal probes capture complementary truth-related features, forming well-defined clusters that reveal the inherent structure of the dataset.

Instruction-following is particularly crucial for large language models (LLMs) to support diverse user requests. While existing work has made progress in aligning LLMs with human preferences, evaluating their capabilities on instruction-following remains a challenge due to complexity and diversity of real-world user instructions. While existing evaluation methods focus on general skills, they suffer from two main shortcomings, i.e., lack of fine-grained task-level evaluation and reliance on singular instruction expression. To address these problems, this paper introduces DINGO, a fine-grained and diverse instruction-following evaluation dataset that has two main advantages: (1) DINGO is based on a manual annotated, fine-grained and multi-level category tree with 130 nodes derived from real-world user requests; (2) DINGO includes diverse instructions, generated by both GPT-4 and human experts. Through extensive experiments, we demonstrate that DINGO can not only provide more challenging and comprehensive evaluation for LLMs, but also provide task-level fine-grained directions to further improve LLMs.

Uncertainty estimation for unlabeled data is crucial to active learning. With a deep neural network employed as the backbone model, the data selection process is highly challenging due to the potential over-confidence of the model inference. Existing methods resort to special learning fashions (e.g. adversarial) or auxiliary models to address this challenge. This tends to result in complex and inefficient pipelines, which would render the methods impractical. In this work, we propose a novel algorithm that leverages noise stability to estimate data uncertainty. The key idea is to measure the output derivation from the original observation when the model parameters are randomly perturbed by noise. We provide theoretical analyses by leveraging the small Gaussian noise theory and demonstrate that our method favors a subset with large and diverse gradients. Our method is generally applicable in various tasks, including computer vision, natural language processing, and structural data analysis. It achieves competitive performance compared against state-of-the-art active learning baselines.

Vehicle-to-Everything (V2X) collaborative perception has recently gained significant attention due to its capability to enhance scene understanding by integrating information from various agents, e.g., vehicles, and infrastructure. However, current works often treat the information from each agent equally, ignoring the inherent domain gap caused by the utilization of different LiDAR sensors of each agent, thus leading to suboptimal performance. In this paper, we propose DI-V2X, that aims to learn Domain-Invariant representations through a new distillation framework to mitigate the domain discrepancy in the context of V2X 3D object detection. DI-V2X comprises three essential components: a domain-mixing instance augmentation (DMA) module, a progressive domain-invariant distillation (PDD) module, and a domain-adaptive fusion (DAF) module. Specifically, DMA builds a domain-mixing 3D instance bank for the teacher and student models during training, resulting in aligned data representation. Next, PDD encourages the student models from different domains to gradually learn a domain-invariant feature representation towards the teacher, where the overlapping regions between agents are employed as guidance to facilitate the distillation process. Furthermore, DAF closes the domain gap between the students by incorporating calibration-aware domain-adaptive attention. Extensive experiments on the challenging DAIR-V2X and V2XSet benchmark datasets demonstrate DI-V2X achieves remarkable performance, outperforming all the previous V2X models. Code is available at https://github.com/Serenos/DI-V2X.

Data-Free Knowledge Distillation (DFKD) enables knowledge transfer from a pretrained teacher to a light-weighted student without original training data. Existing works are limited by a strong assumption that samples used to pretrain the teacher model are balanced, which is, however, unrealistic for many real-world tasks. In this work, we investigated a pragmatic yet under-explored problem: how to perform DFKD from a teacher model pretrained from imbalanced data. We observe a seemingly counter-intuitive phenomenon, i.e., adversarial DFKD algorithms favour minority classes, while causing a disastrous impact on majority classes. We theoretically prove that a biased teacher could cause severe disparity on different groups of synthetic data in adversarial distillation, which further exacerbates the mode collapse of a generator and consequently degenerates the overall accuracy of a distilled student model. To tackle this problem, we propose a class-adaptive regularization method, aiming to encourage impartial representation learning of a generator among different classes under a constrained learning formulation. We devise a primal-dual algorithm to solve the target optimization problem. Through extensive experiments, we show that our method mitigates the biased learning of majority classes in DFKD and improves the overall performance compared with baselines. Code will be available at https://github.com/ldpbuaa/ipad.

The ability to accurately predict the trajectory of surrounding vehicles is a critical hurdle to overcome on the journey to fully autonomous vehicles. To address this challenge, we pioneer a novel behavior-aware trajectory prediction model (BAT) that incorporates insights and findings from traffic psychology, human behavior, and decision-making. Our model consists of behavior-aware, interaction-aware, priority-aware, and position-aware modules that perceive and understand the underlying interactions and account for uncertainty and variability in prediction, enabling higher-level learning and flexibility without rigid categorization of driving behavior. Importantly, this approach eliminates the need for manual labeling in the training process and addresses the challenges of non-continuous behavior labeling and the selection of appropriate time windows. We evaluate BAT's performance across the Next Generation Simulation (NGSIM), Highway Drone (HighD), Roundabout Drone (RounD), and Macao Connected Autonomous Driving (MoCAD) datasets, showcasing its superiority over prevailing state-of-the-art (SOTA) benchmarks in terms of prediction accuracy and efficiency. Remarkably, even when trained on reduced portions of the training data (25%), our model outperforms most of the baselines, demonstrating its robustness and efficiency in predicting vehicle trajectories, and the potential to reduce the amount of data required to train autonomous vehicles, especially in corner cases. In conclusion, the behavior-aware model represents a significant advancement in the development of autonomous vehicles capable of predicting trajectories with the same level of proficiency as human drivers. The project page is available on our GitHub.

Modern GPU clusters inherently exhibit heterogeneity, encompassing various aspects such as computation and communication. This heterogeneity poses a significant challenge for the elastic scheduling of deep learning workloads. Unfortunately, existing elastic schedulers often overlook the impact of heterogeneity on scaling efficiency, resulting in considerably prolonged job completion times.

This paper studies vision-language (V&L) pre-training for deep cross-modal representations. Recently, pre-trained V&L models have shown great success in V&L tasks. However, most existing models apply multi-modal encoders to encode the image and text, at the cost of high training complexity because of the input sequence length. In addition, they suffer from noisy training corpora caused by V&L mismatching. In this work, we propose a lightweight vision-language pre-training (LightVLP) for efficient and effective V&L pre-training. First, we design a new V&L framework with two autoencoders. Each autoencoder involves an encoder, which only takes in unmasked tokens (removes masked ones), as well as a lightweight decoder that reconstructs the masked tokens. Besides, we mask and remove large portions of input tokens to accelerate the training. Moreover, we propose a gated interaction mechanism to cope with noise in aligned image-text pairs. As for a matched image-text pair, the model tends to apply cross-modal representations for reconstructions. By contrast, for an unmatched pair, the model conducts reconstructions mainly using uni-modal representations. Benefiting from the above-mentioned designs, our base model shows competitive results compared to ALBEF while saving 44% FLOPs. Further, we compare our large model with ALBEF under the setting of similar FLOPs on six datasets and show the superiority of LightVLP. In particular, our model achieves 2.2% R@1 gains on COCO Text Retrieval and 1.1% on refCOCO+.

Unprocessed RAW video has shown distinct advantages over sRGB video in video editing and computer vision tasks. However, capturing RAW video is challenging due to limitations in bandwidth and storage. Various methods have been proposed to address similar issues in single image RAW capture through de-rendering. These methods utilize both the metadata and the sRGB image to perform sRGB-to-RAW de-rendering and recover high-quality single-frame RAW data. However, metadata-based methods always require additional computation for online metadata generation, imposing severe burden on mobile camera device for high frame rate RAW video capture. To address this issue, we propose a framework that utilizes frame affinity to achieve high-quality sRGB-to-RAW video reconstruction. Our approach consists of two main steps. The first step, temporal affinity prior extraction, uses motion information between adjacent frames to obtain a reference RAW image. The second step, spatial feature fusion and mapping, learns a pixel-level mapping function using scene-specific and position-specific features provided by the previous frame. Our method can be easily applied to current mobile camera equipment without complicated adaptations or added burden. To demonstrate the effectiveness of our approach, we introduce the first RAW Video De-rendering Benchmark. In this benchmark, our method outperforms state-of-the-art RAW image reconstruction methods, even without image-level metadata.

Pipeline parallelism is a key mechanism to ensure the performance of large model serving systems. These systems need to deal with unpredictable online workloads with low latency and high good put. However, due to the specific characteristics of large models and resource constraints in pipeline parallelism, existing systems struggle to balance resource allocation across pipeline stages. The primary challenge resides in the differential distribution of requests across various stages of the pipeline. We propose QUART, a large model serving system that focuses on optimizing the performance of key stages in pipeline parallelism. QUART dynamically identifies the key stages of the pipeline and introduces an innovative two-level model parameter caching system based on forks to achieve rapid scaling of key stages within seconds. In evaluations with real-world request workloads, QUART reduces average response latency by up to 87.1%) and increases good put by 2.37x compared to the baseline. The experiments demonstrate that QUART effectively reduces tail latency and the average queue length of the pipeline.

Through providing cheap rack and network hosting services, third-party internet data centers (IDCs) have gained significant popularity among cloud service providers. Real-time monitoring of the quality of the IDC network and proactively alarming is crucial to guaranteeing the reliability of cloud services. The prevailing approach to addressing this problem involves utilizing active probes and making evaluations based on the results of single-link or multi-link probing. However, the existing efforts still tend to generate a significant number of unnecessary alerts, resulting in enormous operational costs. For this reason, we first build a large-scale distributed ping-based dial test system that enables monitoring the quality of the IDC network in a many-to-one probe mode. We develop an efficient exporter tool based on the standard Prometheus' data interface to ensure real-time and precise measurement data collection. To quickly and accurately detect potential network issues, we also design a multi-step heuristic-based fault detection and alarm method. Furthermore, we propose a comprehensive alarm life-cycle model based on the results of multi-link probing to guide alarm management in production practice. This system has been successfully deployed in the production environment of Sangfor company's managed cloud for over a year, enabling proactive diagnosis of hundreds of IDC gateway IP addresses. The actual statistical results indicate a significant improvement in the mean time to repair (MTTR) for IDC network failures, reducing it from a few hours to just a few minutes. The average daily number of alarms generated by this system is less than 15, decreasing approximately 85 % compared to before. The alarm accuracy exceeds 95 % and the false negative rate is less than 2 % ■

Federated Learning (FL) enables multiple devices to collaboratively train a shared model while ensuring data privacy. The selection of participating devices in each training round critically affects both the model performance and training efficiency, especially given the vast heterogeneity in training capabilities and data distribution across devices. To deal with these challenges, we introduce a novel device selection solution called FedRank, which is based on an end-to-end, ranking-based model that is pre-trained by imitation learning against state-of-the-art analytical approaches. It not only considers data and system heterogeneity at runtime but also adaptively and efficiently chooses the most suitable clients for model training. Specifically, FedRank views client selection in FL as a ranking problem and employs a pairwise training strategy for the smart selection process. Additionally, an imitation learning-based approach is designed to counteract the cold-start issues often seen in state-of-the-art learning-based approaches. Experimental results reveal that FedRank boosts model accuracy by 5.2% to 56.9%, accelerates the training convergence up to $2.01 \times$ and saves the energy consumption up to 40.1%.

Over the past few years, there has been remarkable progress in research on 3D point clouds and their use in autonomous driving scenarios has become widespread. However, deep learning methods heavily rely on annotated data and often face domain generalization issues. Unlike 2D images whose domains usually pertain to the texture information present in them, the features derived from a 3D point cloud are affected by the distribution of the points. The lack of a 3D domain adaptation benchmark leads to the common practice of training a model on one benchmark (e.g. Waymo) and then assessing it on another dataset (e.g. KITTI). This setting results in two distinct domain gaps: scenarios and sensors, making it difficult to analyze and evaluate the method accurately. To tackle this problem, this paper presents ${\color{Red}\text{LiDAR}}$ Dataset with ${\color{Red}\text{C}}{\text{ross}} - {\color{Red}\text{S}}{\text{ensors}}$ (LiDAR-CS Dataset), which contains large-scale annotated LiDAR point cloud under six groups of different sensors but with the same corresponding scenarios, captured from hybrid realistic LiDAR simulator. To our knowledge, LiDAR-CS Dataset is the first dataset that addresses the sensor-related gaps in the domain of 3D object detection in real traffic. Furthermore, we evaluate and analyze the performance using various baseline detectors and demonstrated its potential applications. Project page: https://opendriving.github.io/lidar-cs.

Vision-aided localization for low-cost mobile robots in diverse environments has attracted widespread attention recently. Although many current systems are applicable in daytime environments, nocturnal visual localization is still an open problem owing to the lack of stable visual information. An insight from most nocturnal scenes is that the static and bright streetlights are reliable visual information for localization. Hence we propose a nocturnal vision-aided localization system in streetlight maps with a novel data association and matching scheme using object detection methods. We leverage the Invariant Extended Kalman Filter (InEKF) to fuse IMU, odometer, and camera measurements for consistent state estimation at night. Furthermore, a tracking recovery module is also designed for tracking failures. Experimental results indicate that our proposed system achieves accurate and robust localization with less than 0.2% relative error of trajectory length in four nocturnal environments.

In the burgeoning field of autonomous vehicles (AVs), trajectory prediction remains a formidable challenge, especially in mixed autonomy environments. Traditional approaches often rely on computational methods such as time-series analysis. Our research diverges significantly by adopting an interdisciplinary approach that integrates principles of human cognition and observational behavior into trajectory prediction models for AVs. We introduce a novel “adaptive visual sector” mechanism that mimics the dynamic allocation of attention human drivers exhibit based on factors like spatial orientation, proximity, and driving speed. Additionally, we develop a “dynamic traffic graph” using Convolutional Neural Networks (CNN) and Graph Attention Networks (GAT) to capture spatio-temporal dependencies among agents. Benchmark tests on the NGSIM, HighD, and MoCAD datasets reveal that our model (GAVA) outperforms state-of-the-art baselines by at least 15.2%, 19.4%, and 12.0%, respectively. Our findings underscore the potential of leveraging human cognition principles to enhance the proficiency and adaptability of trajectory prediction algorithms in AVs.

Pipeline parallelism is an important strategy for improving inference performance in Large Language Models (LLMs). However, we find that different stages of LLM pipelines exhibit distinct performance and request characteristics, posing challenges to system performance in online inference scenarios. To address this issue, we propose Planck, a performance optimization framework tailored for LLM pipeline inference. By balancing request traffic, queue length, and execution time at each stage, Planck introduces a progressive SLO (Service Level Objective) allocation method and a stage instance performance controller. Planck fine-grainedly allocates SLOs to each pipeline stage and dynamically adjusts according to request distribution to control queue length. Through optimizing queue lengths across different stages of the model pipeline, Planck effectively reduces waiting time and tail latency. Evaluations conducted on a real cloud cluster using diverse workloads demonstrate that Planck effectively reduces P99 latency and queue length for each pipeline stage.

The autonomous mapping of large-scale urban scenes presents significant challenges for autonomous robots. To mitigate the challenges, global planning, such as utilizing prior GPS trajectories from OpenStreetMap (OSM), is often used to guide the autonomous navigation of robots for mapping. However, due to factors like complex terrain, unexpected body movement, and sensor noise, the uncertainty of the robot’s pose estimates inevitably increases over time, ultimately leading to the failure of robotic mapping. To address this issue, we propose a novel active loop closure procedure, enabling the robot to actively re-plan the previously planned GPS trajectory. The method can guide the robot to re-visit the previous places where the loop-closure detection can be performed to trigger the back-end optimization, effectively reducing errors and uncertainties in pose estimation. The proposed active loop closure mechanism is implemented and embedded into a real-time OSM-guided robot mapping framework. Empirical results on several large-scale outdoor scenarios demonstrate its effectiveness and promising performance.

Anomaly detection is critical in surveillance systems and patrol robots by identifying anomalous regions in images for early warning. Depending on whether reference data are utilized, anomaly detection can be categorized into anomaly detection with reference and anomaly detection without reference. Currently, anomaly detection without reference, which is closely related to out-of-distribution (OoD) object detection, struggles with learning anomalous patterns due to the difficulty of collecting sufficiently large and diverse anomaly datasets with the inherent rarity and novelty of anomalies. Alternatively, anomaly detection with reference employs the scheme of change detection to identify anomalies by comparing semantic changes between a reference image and a query one. However, there are very few ADr works due to the scarcity of public datasets in this domain. In this paper, we aim to address this gap by introducing the UMAD Benchmark Dataset. To our best knowledge, this is the first benchmark dataset designed specifically for anomaly detection with reference in robotic patrolling scenarios, e.g., where an autonomous robot is employed to detect anomalous objects by comparing a reference and a query video sequences. The reference sequences can be taken by the robot along a specified route when there are no anomalous objects in the scene. The query sequences are captured online by the robot when it is patrolling in the same scene following the same route. Our benchmark dataset is elaborated such that each query image can find a corresponding reference based on accurate robot localization along the same route in the pre-built 3D map, with which the reference and query images can be geometrically aligned using adaptive warping. Besides the proposed benchmark dataset, we evaluate the baseline models of ADr on this dataset. We hope this benchmark dataset will facilitate the advancement of ADr methods in the future. Our UMAD benchmark dataset will be publicly accessible at https://github.com/IMRL/UMAD.

Autonomous cooperative planning (ACP) is a promising technique to improve the efficiency and safety of multi-vehicle interactions for future intelligent transportation systems. However, realizing robust ACP is a challenge due to the aggregation of perception, motion, and communication uncertainties. This paper proposes a novel multi-uncertainty aware ACP (MUACP) framework that simultaneously accounts for multiple types of uncertainties via regularized cooperative model predictive control (RC-MPC). The regularizers and constraints for perception, motion, and communication are constructed according to the confidence levels, weather conditions, and outage probabilities, respectively. The effectiveness of the proposed method is evaluated in the Car Learning to Act (CARLA) simulation platform. Results demonstrate that the proposed MUACP efficiently performs cooperative formation in real time and outperforms other benchmark approaches in various scenarios under imperfect knowledge of the environment.

Ensuring efficient resource allocation while providing service level agreement (SLA) guarantees for end-to-end (E2E) latency is crucial for microservice applications. Although existing studies have made significant contributions towards achieving this objective, they primarily concentrate on static graphs. However, microservice graphs are inherently dynamic during runtime in production environments, necessitating more effective and scalable resource management solutions.In this paper, we present Derm, a new resource management system designed for microservice applications with highly dynamic graphs. Our principal finding is that prioritizing different microservice graphs can lead to a substantial reduction in resource allocation. To take advantage of this opportunity, we develop three main components. The first is a performance model that describes uncertainties of microservice latency through a conditional exponential distribution. The second is a probabilistic quantification of the dynamics of microservice graphs. The third is an optimization method for adjusting the resource allocation of microservices to minimize resource usage. We evaluate Derm in our cluster using real microservice benchmarks and production traces. The results highlight that Derm reduces the resource usage by $68.4 \%$ and lowers SLA violation probability by $6.7 \times$, compared to existing approaches.

Control is a critical module of autonomous driving systems, which ensures safety and enhances the human-machine interface. Due to the diverse control demands dictated by different driving scenarios, autonomous vehicles require a data-intensive, adaptive, and intelligent controller. To speed up the control process and improve the performance in different scenarios, we introduce a novelty federated learning framework FedMG, which efficiently coordinates diverse vehicles to train a collaboratively models while preserving data privacy to tune the control process. Through detailed analysis of driving scenarios, vehicles are clustered to different groups based on driving scenarios to seek a balance between data quality and communication efficiency. It enables the consolidation of several global models, each optimized for peak performance, thereby enhancing the overall system’s effectiveness. Extensive experiments with different numbers of vehicles and a variety of driving scenarios demonstrate the effectiveness of FedMG. The framework significantly reduces cumulative driving errors, achieving reductions ranging from 5.42% to 76.43%, while improving user comfort, with improvements ranging from 2.23% to 34.61% over baselines.

Personalized Federated Learning (PFL) aims to learn a custom model for each distributed client while benefiting from collaborative training in order to overcome the detrimental impact of data heterogeneity. Despite the promising benefits, the existing approaches often compromise the generalization performance of personalized models, as they solely focus on enhancing the personalization capability of models or merely aim to strike a balance between personalization and generalization. Indeed, increasing the personalization capability while preserving the strong generalization performance enabled by collaborative training remains a challenge for PFL, as the two objectives seem to compete with each other. To tackle this challenge, we investigate the relationship between model generalization and personalization under different degrees of heterogeneity. We find that besides the client-specific data distribution, the distribution bias between the unique data distribution of each client and that of the whole population is another critical factor that prominently impacts these two performances. Motivated by the above finding, we propose PFed-DBA, a novel PFL framework that effectively perceives this distribution bias to guide the training process. Concretely, we design the PFL models as a skip-connection network between a shared module for learning the shared representations delivering the common distribution of data across all clients and a personalized module for learning the personalized representations of the heterogeneous distribution bias. Then, we devise corresponding loss functions, aggregation strategy, and updating strategy in order to make the two modules intelligently complement each other. Moreover, we conduct extensive experiments to evaluate the effectiveness of PFed-DBA. The results show that PFed-DBA improves model accuracy to 12.34% at best compared with the state-of-the-art.

Federated Learning (FL) emerges as a new learning paradigm that enables multiple devices to collaboratively train a shared model while preserving data privacy. However, intensive memory footprint during the training process severely bottlenecks the deployment of FL on resource-limited mobile devices in real-world cases. Thus, a framework that can effectively reduce the memory footprint while guaranteeing training efficiency and model accuracy is crucial for FL.In this paper, we propose SmartFreeze, a framework that effectively reduces the memory footprint by conducting the training in a progressive manner. Instead of updating the full model in each training round, SmartFreeze divides the shared model into blocks consisting of a specified number of layers. It first trains the front block with a well-designed output module, safely freezes it after convergence, and then triggers the training of the next one. This process iterates until the whole model has been successfully trained. In this way, the backward computation of the frozen blocks and the corresponding memory space for storing the intermediate outputs and gradients are effectively saved. Except for the progressive training framework, SmartFreeze consists of the following two core components: a pace controller and a participant selector. The pace controller is designed to effectively monitor the training progress of each block at runtime and safely freezes them after convergence while the participant selector selects the right devices to participate in the training for each block by jointly considering the memory capacity, the statistical and system heterogeneity. Extensive experiments are conducted to evaluate the effectiveness of SmartFreeze on both simulation and hardware testbeds. The results demonstrate that SmartFreeze effectively reduces average memory usage by up to 82%. Moreover, it simultaneously improves the model accuracy by up to 83.1% and accelerates the training process up to 2.02 ×.

While RAW images are efficient for image editing and perception tasks, their large size can strain camera storage and bandwidth. Reconstruction methods of RAW images from sRGB data typically require additional metadata from the RAW image, which increases camera processing computations. To address this problem, we propose using Prior Meta as a reference to reconstruct the RAW data instead of relying on per-image metadata. Prior metadata is extracted offline from reference RAW images, which are usually part of the training dataset and have similar scenes and light conditions as the target image. With this prior metadata, the camera does not need to provide any extra processing other than the sRGB images, and our model can autonomously find the desired prior information. To achieve this, we design a three-step pipeline. First, we build a pixel searching network that can find the most similar pixels in the reference RAW images as prior information. Then, in the second step, we compress the large-scale reference images to about 0.02% of their original size to reduce the searching cost. Finally, in the last step, we develop a neural network reconstructor to reconstruct the high-fidelity RAW images. Our model achieves comparable, and even better, performance than RAW reconstruction methods based on metadata.

As autonomous driving systems increasingly become part of daily transportation, the ability to accurately anticipate and mitigate potential traffic accidents is paramount. Traditional accident anticipation models primarily utilizing dashcam videos are adept at predicting when an accident may occur but fall short in localizing the incident and identifying involved entities. Addressing this gap, this study introduces a novel framework that integrates Large Language Models (LLMs) to enhance predictive capabilities across multiple dimensions-what, when, and where accidents might occur. We develop an innovative chain-based attention mechanism that dynamically adjusts to prioritize high-risk elements within complex driving scenes. This mechanism is complemented by a three-stage model that processes outputs from smaller models into detailed multimodal inputs for LLMs, thus enabling a more nuanced understanding of traffic dynamics. Empirical validation on the DAD, CCD, and A3D datasets demonstrates superior performance in Average Precision (AP) and Mean Time-To-Accident (mTTA), establishing new benchmarks for accident prediction technology. Our approach not only advances the technological framework for autonomous driving safety but also enhances human-AI interaction, making predictive insights generated by autonomous systems more intuitive and actionable.

Accurately and promptly predicting accidents among surrounding traffic agents from camera footage is crucial for the safety of autonomous vehicles (AVs). This task presents substantial challenges stemming from the unpredictable nature of traffic accidents, their long-tail distribution, the intricacies of traffic scene dynamics, and the inherently constrained field of vision of onboard cameras. To address these challenges, this study introduces a novel accident anticipation framework for AVs, termed CRASH. It seamlessly integrates five components: object detector, feature extractor, object-aware module, context-aware module, and multi-layer fusion. Specifically, we develop the object-aware module to prioritize high-risk objects in complex and ambiguous environments by calculating the spatial-temporal relationships between traffic agents. In parallel, the context-aware is also devised to extend global visual information from the temporal to the frequency domain using the Fast Fourier Transform (FFT) and capture fine-grained visual features of potential objects and broader context cues within traffic scenes. To capture a wider range of visual cues, we further propose a multi-layer fusion that dynamically computes the temporal dependencies between different scenes and iteratively updates the correlations between different visual features for accurate and timely accident prediction. Evaluated on real-world datasets-Dashcam Accident Dataset (DAD), Car Crash Dataset (CCD), and AnAn Accident Detection (A3D) datasets-our model surpasses existing top baselines in critical evaluation metrics like Average Precision (AP) and mean Time-To-Accident (mTTA). Importantly, its robustness and adaptability are particularly evident in challenging driving scenarios with missing or limited training data, demonstrating significant potential for application in real-world autonomous driving systems.

Adapting Large Language Models (LLMs) to new tasks through fine-tuning has been made more efficient by the introduction of Parameter-Efficient Fine-Tuning (PEFT) techniques, such as LoRA. However, these methods often underperform compared to full fine-tuning, particularly in scenarios involving complex datasets. This issue becomes even more pronounced in complex domains, highlighting the need for improved PEFT approaches that can achieve better performance. Through a series of experiments, we have uncovered two critical insights that shed light on the training and parameter inefficiency of LoRA. Building on these insights, we have developed HydraLoRA, a LoRA framework with an asymmetric structure that eliminates the need for domain expertise. Our experiments demonstrate that HydraLoRA outperforms other PEFT approaches, even those that rely on domain knowledge during the training and inference phases. Our anonymous codes are submitted with the paper and will be publicly available. Code is available: https://github.com/Clin0212/HydraLoRA.

The deployment of ML serving applications, featuring multiple inference functions on serverless platforms, has gained substantial popularity, leading to numerous developments of new systems. However, these systems often focus on optimizing resource provisioning and cold start management separately, ultimately resulting in higher monetary costs. This paper introduces SMIless, a highly efficient serverless system tailored for serving DAG-based ML inference in heterogeneous environments. SMIless effectively co-optimizes resource configuration and cold-start management in the context of dynamic invocations. This is achieved by seamlessly integrating adaptive pre-warming windows, striking an effective balance between performance and cost. We have implemented SMIless on top of OpenFaaS and conducted extensive evaluations using real-world ML serving applications. The experimental results demonstrate that SMIless can achieve up to a $5.73 \times$ reduction in the overall costs while meeting the SLA requirements for all user requests, surpassing the performance of state-of-the-art solutions.

Federated Learning (FL) emerges as a new learning paradigm that enables multiple devices to collaboratively train a shared model while preserving data privacy. However, one fundamental and prevailing challenge that hinders the deployment of FL on mobile devices is the memory limitation. This paper proposes FedHybrid, a novel framework that effectively reduces the memory footprint during the training process while guaranteeing the model accuracy and the overall training progress. Specifically, FedHybrid first selects the participating devices for each training round by jointly evaluating their memory budget, computing capability, and data diversity. After that, it judiciously analyzes the computational graph and generates an execution plan for each selected client in order to meet the corresponding memory budget while minimizing the training delay through employing a hybrid of recomputation and compression techniques according to the characteristic of each tensor. During the local training process, FedHybrid carries out the execution plan with a well-designed activation compression technique to effectively achieve memory reduction with minimum accuracy loss. We conduct extensive experiments to evaluate FedHybrid on both simulation and off-the-shelf mobile devices. The experiment results demonstrate that FedHybrid achieves up to a 39.1% increase in model accuracy and a 15.5X reduction in wall clock time under various memory budgets compared with the baselines.

A significant proportion of energy consumed in modern data centers and clouds is dedicated to provisioning idle servers for maintaining Quality of Service guarantees. Various studies have been conducted exploring dynamic provisioning in data centers with the objective of reducing overall energy consumption. However, many of these studies assume a fixed energy cost per operating server where each server can only handle one job within a given time slot. In this paper, we address a new and practical problem that involves speed scaling of multiple servers within a data center. Specifically, we consider a scenario where each server can handle multiple jobs simultaneously, and the energy consumed is a piece-wise convex function that depends on processing speed. In addition, turning on a server incurs a substantial energy cost.

As an emerging technology, Digital Twin (DT) can provide a virtual presentation of the physical Intelligent Trans-portation Systems (ITS) to enhance the applications of ITS such as cooperation perception. In cooperative perception, accurate location is crucial for selecting proper cooperative vehicles (CoVs) to improve the perception performance. However, due to the high mobility of vehicles, the deviation between DT and physical world may lead to non-negligible location errors, which raises the challenges for achieving efficient CoV selection in cooperative perception. In this paper, we propose a DT-empowered multi-vehicle cooperative perception system, in which the CoV selection and bandwidth allocation are jointly optimized to improve the performance of cooperative perception. Specifically, an asyn-chronous federated learning scheme is deployed in DT for location prediction to mitigate the effect of the deviation. Based on the prediction results, the problem of joint predictive scheduling and bandwidth allocation is then formulated as the average delay minimization problem while reaching the required performances. The adaptive CoV selection and bandwidth allocation algorithm based on deep reinforcement learning is proposed to find the optimal scheduling strategy. Simulation results demonstrate that the proposed algorithm achieves the lowest average delay while effectively guaranteeing the performance requirement.

Summary Real‐time estimation of crowd counting in underground metro systems, constrained by limited space, is crucial for managing heightened pedestrian volumes and responding promptly to emergencies. To address this challenge, we propose a passenger state transition‐based model, called STRmt, designed for the seamless and continuous monitoring of real‐time crowd movement within service areas of stations and trains, leveraging auto fare collection systems (AFC) as a comprehensive sensor network. Our innovation lies in modeling the dynamic movement of passengers within a metro system over time as a state transition process aligned with the train schedule. To achieve this, we introduce a spatio‐temporal deep learning framework, denoted as STnet, designed to dynamically predict these state transitions. The performance of our method is rigorously assessed through extensive experiments conducted spanning 2 years in Shenzhen, China, utilizing AFC data, train schedule data, and weather data. The results demonstrate that the proposed method surpasses baseline methods, achieving an estimation precision of 0.92.

A hybrid active-passive wireless sensor network (HWSN) is a cost-effective and energy-efficient way for localization systems. The active sensor nodes, which are targets, can locate themselves by sending wireless power signals to power up the passive sensors, or via cooperative localization among targets. The common used power allocation approaches e.g., semi-definite programming (SDP), are aiming to minimize the overall localization error with energy constraints. However, the global optimum may lead to issues such as decreased individual performance and unbalanced network resources. To address the above challenges, we propose a game theory power allocation framework for the cooperative localization of HWSNs in this paper. We present the Fisher information matrix (FIM) and the corresponding squared position error bound (SPEB) for two types of HWSNs, which are the general cooperation network and tree topology network. For the general cooperation network, we derive the closed-form solution of optimal power allocation using the general bargaining equilibrium. In the tree topology network, we propose a Stackelberg equilibrium approach that effectively adapts the power allocation scheme based on the obtained hierarchical information structure. Extensive simulation results demonstrate that the proposed method outperforms semidefinite programming (SDP) and equal power allocation scheme (EQ) in position estimation error. Specifically, compared to other schemes, our proposed scheme has improved the positioning accuracy by 75% in the general cooperation network and by 70% in the tree topology network.

Characterizing the sensing and communication performance tradeoff in integrated sensing and communication (ISAC) systems is challenging in the applications of learning-based human motion recognition. This is because of the large experimental data sets and the black-box nature of deep neural networks. This article presents SDP3, a Simulation-Driven Performance Predictor and oPtimizer, which consists of SDP3 data simulator, SDP3 performance predictor and SDP3 performance optimizer. Specifically, the SDP3 data simulator generates vivid wireless sensing data sets in a virtual environment, the SDP3 performance predictor predicts the sensing performance based on the curve fitting method, and the SDP3 performance optimizer investigates the sensing and communication performance tradeoff analytically. It is shown that the simulated sensing data set matches the experimental data set very well in the motion recognition accuracy. By leveraging SDP3, it is found that the achievable region of recognition accuracy and communication throughput consists of a communication saturation zone, a sensing saturation zone, and a communication-sensing adversarial zone, of which the desired balanced performance for ISAC systems lies in the third one.

Human activity recognition (HAR) based on WiFi channel state information (CSI) has received a lot of attentions recently due to its nonintrusive nature. Most CSI-based HAR systems use a WiFi router and a computing terminal for centralized processing, which makes it difficult to achieve real-time wide-range recognition. Recently, lightweight Artificial Intelligence Internet of Things (AIoT) devices are widely deployed. The equipped WiFi chips within such devices can collect and process CSI data in a distributed way. Thus, the AIoT devices extend the detection range of collecting CSI and enrich the applications. However, the memories of the AIoT devices are constrained and lack of appropriate lightweight CSI processing strategies. To address these challenges, we propose the LiWi-HAR system which employs a comprehensive lightweight CSI processing strategy in WiFi-based AIoT devices. The proposed lightweight CSI processing strategy extracts the main related features while compressing the data size. Then, a double hidden layer BP neural network based on particle swarm optimization (PSO-BPNN) algorithm is developed for HAR. In this case, the computing memory occupation of the device is effectively reduced, and the real-time high-accurate recognition is achieved. Extensive experimental results present that the efficiency of our system significantly outperforms other centralized deep learning-based systems and the recognition accuracy achieves 91.7%.

The recent advancements in integrated sensing and communications (ISACs) technology have introduced new possibilities to address the quality of communication and high-resolution positioning requirements in the next-generation wireless communication network (6G) vehicle-to-everything (V2X). Simultaneously providing high-accurate positioning and high-communication capacity (CC) for the intelligent service of the vehicle target is challenging. In this article, we propose a reconfigurable intelligent surface (RIS)-assisted 6G V2X system to achieve highly accurate positioning of the vehicle target with basic communication requirements. We provide the CC and the 3-D fisher information matrix (FIM) formulations of the vehicle target. We demonstrate the direct impact of phase modulation in the reflector units on joint positioning accuracy and CC performance. Meanwhile, we design a flexible deep deterministic policy gradient (FL-DDPG) algorithm network with an$\epsilon $-greedy strategy to solve the high-dimensional nonconvex optimization problem, achieves minimal positioning error while satisfying various CC requirements. Simulation results demonstrate that the FL-DDPG algorithm enhances positioning accuracy by a minimum of 89% and improves the achievable rate of the vehicle target by nearly 3 times, which outperforms traditional mathematical methods. Compared with classical deep reinforcement learning methods, FL-DDPG achieves better positioning accuracy while satisfying the communication requirements. When confronting imperfect channel, FL-DDPG enables addressing the channel estimation errors effectively on the ISAC system.

For future 6G systems, reconfigurable intelligent surface (RIS) controls phase shift of the reflective unit to improve the channel, which affects the received signal strength (RSS) of the microwave. In this paper, we introduce the RIS as anchors into the cooperative localization wireless sensor network (RACLN) to locate the sensor nodes. The RIS in RACLN can be a base station that provides multiple passive antennas for highly accurate localization. We derive the Cramér-Rao lower bound (CRLB) and the related squared position error bound (SPEB) for RACLN. The formulations indicate that the phase shift control of the RIS can improve the localization accuracy effectively. However, determining the appropriate phase shifts poses a nonlinear, non-convex integer programming problem. Thus, we propose a semi-definite programming (SDP) based phase optimization algorithm (SDP-PO) by relaxing the objective and achieving the minimum SPEB. Further, we also develop a low-complexity phase optimization (LC-PO) algorithm to reduce the dimension of the phase shift vector of SDP-PO. The simulation results demonstrate that the SPEB of RACLN is 94:68% smaller than the WSN using sensor nodes as anchors. In addition, The proposed SDP-PO algorithms outperform the genetic algorithm (GA) and the alternate optimization (AO) with 17.12% and 37.49%.

Localization is essential for network management of the wireless backscattered sensor networks (WBSNs). In large-scale WBSN, the cooperative localization among the passive nodes effectively improves the localization accuracy. Meanwhile, reconfigurable intelligent surface (RIS) motivates nodes to gain better spatial channel using passive beamforming or phase modulation. In this article, we analyze the impact of passive beamforming for RIS on the localization accuracy of cooperative localization in the WBSN system. We derive the Fisher information matrix (FIM) and the spatial position error bound for the fully connected communication network system. We demonstrate that the phase modulation of RIS reflection units affect the localization accuracy of the cooperative localization WBSN system. However, RIS passive beamforming as a discrete and nonconvex integer programming problem is difficult to solve. Then, we propose a Monte Carlo-based random RIS passive beamforming to achieve the maximum localization accuracy. We apply Gibbs sampling and resampling methods to generate the phase shift vector samples of RIS. The sample with the highest localization accuracy is considered as the optimal solution. The simulation results demonstrate that our proposed method for RIS passive beamforming can improve 34.5% localization accuracy in the Line-of-Sight (LoS) case, while the genetic algorithm (GA) is 6.8%. In the Non-LoS (NLoS) environment, the localization accuracy improvement of our proposed method reaches 97%, and GA can only reach 85% as the comparison.

Abstract Fine-tuning deep neural networks pre-trained on large scale datasets is one of the most practical transfer learning paradigm given limited quantity of training samples. To obtain better generalization, using the starting point as the reference (SPAR), either through weights or features, has been successfully applied to transfer learning as a regularizer. However, due to the domain discrepancy between the source and target task, there exists obvious risk of negative transfer in a straightforward manner of knowledge preserving. In this paper, we propose a novel transfer learning algorithm, introducing the idea of Target-awareness REpresentation Disentanglement ( $$\textrm{TRED}$$ TRED ), where the relevant knowledge with respect to the target task is disentangled from the original source model and used as a regularizer during fine-tuning the target model. Two alternative approaches, maximizing Maximum Mean Discrepancy (Max-MMD) and minimizing mutual information (Min-MI) are introduced to achieve the desired disentanglement. Experiments on various real world datasets show that our method stably improves the standard fine-tuning by more than 2% in average. $$\textrm{TRED}$$ TRED also outperforms related state-of-the-art transfer learning regularizers such as $$\mathrm {L^2\text {-}SP}$$ L 2 - SP , $$\textrm{AT}$$ AT , $$\textrm{DELTA}$$ DELTA , and $$\textrm{BSS}$$ BSS . Moreover, our solution is compatible with different choices of disentangling strategies. While the combination of Max-MMD and Min-MI typically achieves higher accuracy, only using Max-MMD can be a preferred choice in applications with low resource budgets.

Fusing features from different sources is a critical aspect of many computer vision tasks. Existing approaches can be roughly categorized as parameter-free or learnable operations. However, parameter-free modules are limited in their ability to benefit from offline learning, leading to poor performance in some challenging situations. Learnable fusing methods are often space-consuming and time-consuming, particularly when fusing features with different shapes. To address these shortcomings, we conducted an in-depth analysis of the limitations associated with both fusion methods. Based on our findings, we propose a generalized module named Asymmetric Convolution Module (ACM). This module can learn to encode effective priors during offline training and efficiently fuse feature maps with different shapes in specific tasks. Specifically, we propose a mathematically equivalent method for replacing costly convolutions on concatenated features. This method can be widely applied to fuse feature maps across different shapes. Furthermore, distinguished from parameter-free operations that can only fuse two features of the same type, our ACM is general, flexible, and can fuse multiple features of different types. To demonstrate the generality and efficiency of ACM, we integrate it into several state-of-the-art models on three representative vision tasks. Extensive experimental results on three tasks and several datasets demonstrate that our new module can bring significant improvements and noteworthy efficiency.

Although face swapping has attracted much attention in recent years, it remains a challenging problem. Existing methods leverage a large number of data samples to explore the intrinsic properties of face swapping without considering the semantic information of face images. Moreover, the representation of the identity information tends to be fixed, leading to suboptimal face swapping. In this paper, we present a simple yet efficient method named FaceSwapper, for one-shot face swapping based on Generative Adversarial Networks. Our method consists of a disentangled representation module and a semantic-guided fusion module. The disentangled representation module comprises an attribute encoder and an identity encoder, which aims to achieve the disentanglement of the identity and attribute information. The identity encoder is more flexible, and the attribute encoder contains more attribute details than its competitors. Benefiting from the disentangled representation, FaceSwapper can swap face images progressively. In addition, semantic information is introduced into the semantic-guided fusion module to control the swapped region and model the pose and expression more accurately. Experimental results show that our method achieves state-of-the-art results on benchmark datasets with fewer training samples.

Deep convolutional neural networks (CNNs) can be easily tricked to give incorrect outputs by adding tiny perturbations to the input that are imperceptible to humans. This makes them susceptible to adversarial attacks, and poses significant security risks to deep learning systems, and presents a great challenge in making CNNs robust against such attacks. An influx of defense strategies have thus been proposed to improve the robustness of CNNs. Current attack methods, however, may fail to accurately or efficiently evaluate the robustness of defending models. In this paper, we thus propose a unified$ \ell _{p}$white-box attack strategy, LAFIT, to harness the defender's latent features in its gradient descent steps, and further employ a new loss function to normalize logits to overcome floating-point-based gradient masking. We show that not only is it more efficient, but it is also a stronger adversary than the current state-of-the-art when examined across a wide range of defense mechanisms. This suggests that adversarial attacks/defenses could be contingent on the effective use of the defender's hidden components, and robustness evaluation should no longer view models holistically.

Short-term OD flow (i.e. the number of passenger traveling between stations) prediction is crucial to traffic management in metro systems. The delayed effect in latest complete OD flow collection and complex spatiotemporal correlations of OD flows in high dimension make it challengeable to predict short-term OD flow. Existing methods need to be improved due to not fully utilizing the real-time passenger mobility data and not sufficiently modeling the implicit correlation of the mobility patterns between stations. In this paper, we propose a Completion based Adaptive Heterogeneous Graph Convolution Spatiotemporal Predictor. The novelty is mainly reflected in two aspects. The first is to model real-time mobility evolution by establishing the implicit correlation between observed OD flows and the prediction target OD flows in high dimension based on a key data-driven insight: the destination distributions of the passengers departing from a station are correlated with other stations sharing similar attributes (e.g. geographical location, region function). The second is to complete the latest incomplete OD flows by estimating the destination distribution of unfinished trips through considering the real-time mobility evolution and the time cost between stations, which is the base of time series prediction and can improve the model’s dynamic adaptability. Extensive experiments on two real world metro datasets demonstrate the superiority of our model over other competitors with the biggest model performance improvement being nearly 4%. In addition, the data complete framework we propose can be integrated into other models to improve their performance up to 2.1%.

Among various value decomposition-based multiagent reinforcement learning (MARL) algorithms, the overall performance of the multiagent system is represented by a scalar global Q value and optimized by minimizing the temporal difference (TD) error with respect to that global Q value. However, the global Q value cannot accurately model the distributed dynamics of the multiagent system, since it is only a simplified representation for different individual Q values of agents. To explicitly consider the correlations between different cooperative agents, in this article, we propose a distributional framework and construct a practical model called distributional multiagent cooperation (DMAC) from a novel distributional perspective. Specifically, in DMAC, we view the individual Q value for the executed action of a random agent as a value distribution, whose expectation can further represent the overall performance. Then, we employ distributional RL to minimize the difference between the estimated distribution and its target for the optimization. The advantage of DMAC is that the distributed dynamics of agents can be explicitly modeled, and this results in better performance. To verify the effectiveness of DMAC, we conduct extensive experiments under nine different scenarios of the StarCraft Multiagent Challenge (SMAC). Experimental results show that the DMAC can significantly outperform the baselines with respect to the average median test win rate.

This article presents a novel efficient experience-replay-based adaptive dynamic programming (ADP) for the optimal control problem of a class of nonlinear dynamical systems within the Hamiltonian-driven framework. The quasi-Hamiltonian is presented for the policy evaluation problem with an admissible policy. With the quasi-Hamiltonian, a novel composite critic learning mechanism is developed to combine the instantaneous data with the historical data. In addition, the pseudo-Hamiltonian is defined to deal with the performance optimization problem. Based on the pseudo-Hamiltonian, the conventional Hamilton-Jacobi-Bellman (HJB) equation can be represented in a filtered form, which can be implemented online. Theoretical analysis is investigated in terms of the convergence of the adaptive critic design and the stability of the closed-loop systems, where parameter convergence can be achieved under a weakened excitation condition. Simulation studies are investigated to verify the efficacy of the presented design scheme.

A common approach to improving resource utilization in data centers is to adaptively provision resources based on the actual workload. One fundamental challenge of doing this in microservice management frameworks, however, is that different components of a service can exhibit significant differences in their impact on end-to-end performance. To make resource management more challenging, a single microservice can be shared by multiple online services that have diverse workload patterns and SLA requirements. We present an efficient resource management system, namely Erms, for guaranteeing SLAs with high probability in shared microservice environments. Erms profiles microservice latency as a piece-wise linear function of the workload, resource usage, and interference. Based on this profiling, Erms builds resource scaling models to optimally determine latency targets for microservices with complex dependencies. Erms also designs new scheduling policies at shared microservices to further enhance resource efficiency. Experiments across microservice benchmarks as well as trace-driven simulations demonstrate that Erms can reduce SLA violation probability by 5× and more importantly, lead to a reduction in resource usage by 1.6×, compared to state-of-the-art approaches.

As Internet of Things devices become widely used in the real-world, it is crucial to protect them from malicious intrusions. However, the data scarcity of IoT limits the applicability of traditional intrusion detection methods, which are highly data-dependent. To address this, in this article, we propose the Open-Set Dandelion Network (OSDN) based on unsupervised heterogeneous domain adaptation in an open-set manner. The OSDN model performs intrusion knowledge transfer from the knowledge-rich source network intrusion domain to facilitate more accurate intrusion detection for the data-scarce target IoT intrusion domain. Under the open-set setting, it can also detect newly-emerged target domain intrusions that are not observed in the source domain. To achieve this, the OSDN model forms the source domain into a dandelion-like feature space in which each intrusion category is compactly grouped and different intrusion categories are separated, i.e., simultaneously emphasising inter-category separability and intra-category compactness. The dandelion-based target membership mechanism then forms the target dandelion. Then, the dandelion angular separation mechanism achieves better inter-category separability, and the dandelion embedding alignment mechanism further aligns both dandelions in a finer manner. To promote intra-category compactness, the discriminating sampled dandelion mechanism is used. Assisted by the intrusion classifier trained using both known and generated unknown intrusion knowledge, a semantic dandelion correction mechanism emphasises easily-confused categories and guides better inter-category separability. Holistically, these mechanisms form the OSDN model that effectively performs intrusion knowledge transfer to benefit IoT intrusion detection. Comprehensive experiments on several intrusion datasets verify the effectiveness of the OSDN model, outperforming three state-of-the-art baseline methods by 16.9%. The contribution of each OSDN constituting component, the stability and the efficiency of the OSDN model are also verified.

As the applications of AI proliferate, it is critical to increase the throughput of online DNN inference services. Multi-process service (MPS) improves the utilization rate of GPU resources by spatial-sharing, but it also brings unique challenges. First, interference between co-located DNN models deployed on the same GPU must be accurately modeled. Second, inference tasks arrive dynamically online, and each task needs to be served within a bounded time to meet the service-level objective (SLO). Third, the problem of fragments has become more serious. To address the above three challenges, we propose anIntelligentScheduling orchestrator for multi-GPU inference servers with spatio-temporalSharing (InSS), aiming to maximize the system throughput.InSSexploits two key innovations: i) An interference-aware latency analytical model which estimates the task latency. ii) A two-stage intelligent scheduler is tailored to jointly optimize the model placement, GPU resource allocation and adaptively decides batch size by coupling the latency analytical model. Our prototype implementation on four NVIDIA A100 GPUs shows thatInSScan improve the throughput by up to 86% compared to the state-of-the-art GPU schedulers, while satisfying SLOs. We further show the scalability ofInSSon 64 GPUs.

Federated Learning (FL) enables multiple devices to collaboratively train a shared model while preserving data privacy. Ever-increasing model complexity coupled with limited memory resources on the participating devices severely bottlenecks the deployment of FL in real-world scenarios. Thus, a framework that can effectively break the memory wall while jointly taking into account the hardware and statistical heterogeneity in FL is urgently required. In this article, we proposeSmartSplita framework that effectively reduces the memory footprint on the device side while guaranteeing the training progress and model accuracy for heterogeneous FL through model splitting. Towards this end,SmartSplitemploys a hierarchical structure to adaptively guide the overall training process. In each training round, the central manager, hosted on the server, dynamically selects the participating devices and sets the cutting layer by jointly considering the memory budget, training capacity, and data distribution of each device. The MEC manager, deployed within the edge server, proceeds to split the local model and perform training of the server-side portion. Meanwhile, it fine-tunes the splitting points based on the time-evolving statistical importance. The on-device manager, embedded inside each mobile device, continuously monitors the local training status while employing cost-aware checkpointing to match the runtime dynamic memory budget. Extensive experiments on representative datasets are conducted on both commercial off-the-shelf mobile device testbeds. The experimental results show thatSmartSplitexcels in FL training on highly memory-constrained mobile SoCs, offering up to a 94% peak latency reduction and 100-fold memory savings. It enhances accuracy performance by 1.49%-57.18% and adaptively adjusts to dynamic memory budgets through cost-aware recomputation

Deploying federated learning (FL) in edge clouds poses challenges, especially when multiple models are concurrently trained in resource-constrained edge environments. Existing research on federated edge learning has predominantly focused on client selection for training a single FL model, typically with a fixed learning topology. Preliminary experiments indicate that FL models with adaptable topologies exhibit lower learning costs compared to those with fixed topologies. This paper delves into the intricacies of jointly selecting participants and learning topologies for multiple FL models simultaneously trained in the edge cloud. The problem is formulated as an integer non-linear programming problem, aiming to minimize total learning costs associated with all FL models while adhering to edge resource constraints. To tackle this challenging optimization problem, we introduce a two-stage algorithm that decouples the original problem into two sub-problems and iteratively addresses them separately with efficient heuristics. Our method enhances resource competition and load balancing in edge clouds by allowing FL models to choose participants and learning topologies independently. Extensive experiments conducted with real-world networks and FL datasets affirm the better performance of our algorithm, demonstrating lower average total costs with up to 33.5% and 39.6% compared to previous methods designed for multi-model FL.

Microservices have transformed monolithic applications into lightweight, self-contained, and isolated application components, establishing themselves as a dominant paradigm for application development and deployment in public clouds such as Google and Alibaba. Autoscaling emerges as an efficient strategy for managing resources allocated to microservices’ replicas. However, the dynamic and intricate dependencies within microservice chains present challenges to the effective management of scaled microservices. Additionally, the centralized autoscaling approach can encounter scalability issues, especially in the management of large-scale microservice-based clusters. To address these challenges and enhance scalability, we propose an innovative distributed resource provisioning approach for microservices based on the Twin Delayed Deep Deterministic Policy Gradient algorithm. This approach enables effective autoscaling decisions and decentralizes responsibilities from a central node to distributed nodes. Comparative results with state-of-the-art approaches, obtained from a realistic testbed and traces, indicate that our approach reduces the average response time by 15% and the number of failed requests by 24%, validating improved scalability as the number of requests increases.

Mobile edge computing emerges to serve mobile users with low-latency computation offloading in edge networks, which are resource-constrained with massive users and workloads. However, existing communication and computing resource allocation schemes for offloaded tasks aren't efficient enough, where finished tasks still occupy resources, wasting constrained resources. Besides, the multi-user offloading is usually for scenarios of one task per user, ignoring real-worldmulti-taskoffloading scenarios where each user has multiple tasks, lack generality and flexibility. Meanwhile, local computing resource allocation schemes in multi-task scenarios ignore resource readjustment, causing low resource utilization. To solve these problems, we propose ECO-GAME, an efficient multi-task offloading scheme, which dynamically allocates bandwidth and computing resources to unfinished tasks, resulting in high resource utilization. We initially formulate the multi-task offloading problem as the game minimizing each user's cost, which is NP-hard. Thus we re-formulate the game utilizing potential games to optimize user's objective either locally or globally, and prove the existence of its Nash equilibrium. We then design an efficient multi-task offloading algorithm to obtain an approximate solution in polynomial time, together with computational complexity analysis. We further conduct performance evaluation on ECO-GAME utilizing price of anarchy. Numerical results demonstrate the efficiency of ECO-GAME, and show ECO-GAME reduces 49.2% cost over the state-of-the-art work, and scales well with the increasing number of tasks and users.

A common approach to improving resource utilization in data centers is to adaptively provision resources based on the actual workload. One fundamental challenge of doing this in microservice management frameworks, however, is that different components of a service can exhibit significant differences in their impact on end-to-end performance. To make resource management more challenging, a single microservice can be shared by multiple online services that have diverse workload patterns and SLA requirements.

Owing to the non-i.i.d. nature of client data, channel neurons in federated-learned models may specialize to distinct features for different clients. Yet, existing channel-sparse federated learning (FL) algorithms prescribe fixed sparsity strategies for client models, and may thus prevent clients from training channel neurons collaboratively. To minimize the impact of sparsity on FL convergence, we propose Flado to improve the alignment of client model update trajectories by tailoring the sparsities of individual neurons in each client. Empirical results show that while other sparse methods are surprisingly impactful to convergence, Flado can not only attain the highest task accuracies with unlimited budget across a range of datasets, but also significantly reduce the amount of floating-point operations (FLOPs) required for training more than by 10× under the same communications budget, and push the Pareto frontier of communication/computation trade-off notably further than competing FL algorithms.

Monocular 3D object detection has become a mainstream approach in automatic driving for its easy application. A prominent advantage is that it does not need Li-DAR point clouds during the inference. However, most current methods still rely on 3D point cloud data for labeling the ground truths used in the training phase. This inconsistency between the training and inference makes it hard to utilize the large-scale feedback data and increases the data collection expenses. To bridge this gap, we propose a new weakly supervised monocular 3D objection detection method, which can train the model with only 2D labels marked on images. To be specific, we explore three types of consistency in this task, i.e. the projection, multi-view and direction consistency, and design a weakly-supervised architecture based on these consistencies. Moreover, we propose a new 2D direction labeling method in this task to guide the model for accurate rotation direction prediction. Experiments show that our weakly-supervised method achieves comparable performance with some fully supervised methods. When used as a pre-training method, our model can significantly outperform the corresponding fully-supervised baseline with only 1/3 3D labels.

Attackers can deceive neural networks by adding human imperceptive perturbations to their input data; this reveals the vulnerability and weak robustness of current deep-learning networks. Many attack techniques have been proposed to evaluate the model's robustness. Gradient-based attacks suffer from severely overestimating the robustness. This paper identifies that the relative error in calculated gradients caused by floating-point errors, including floating-point underflow and rounding errors, is a fundamental reason why gradient-based attacks fail to accurately assess the model's robustness. Although it is hard to eliminate the relative error in the gradients, we can control its effect on the gradient-based attacks. Correspondingly, we propose an efficient loss function by minimizing the detrimental impact of the floating-point errors on the attacks. Experimental results show that it is more efficient and reliable than other loss functions when examined across a wide range of defence mechanisms.

Federated Learning coordinates multiple clients to collaboratively train a shared model while preserving data privacy. However, the training data with noisy labels located on the participating clients severely harm the model performance. In this paper, we propose FedCoop, a cooperative Federated Learning framework for noisy labels. FedCoop mainly contains three components and conducts robust training in two phases, data selection and model training. In the data selection phase, in order to mitigate the confirmation bias caused by a single client, the Loss Transformer intelligently estimates the probability of each sample’s label to be clean through cooperating with the helper clients, which have high data trustability and similarity. After that, the Feature Comparator evaluates the label quality for each sample in terms of latent feature space in order to further improve the robustness of noisy label detection. In the model training phase, the Feature Matcher trains the model on both the noisy and clean data in a semi-supervised manner to fully utilize the training data and exploits the feature of global class to increase the consistency of pseudo labeling across the clients. The experimental results show FedCoop outperforms the baselines on various datasets with different noise settings. It effectively improves the model accuracy up to 62% and 27% on average compared with the baselines.

To fully utilize computing resources, cloud providers such as Google and Alibaba choose to co-locate online services with batch processing applications in their data centers. By implementing unified resource management policies, different types of complex computing jobs request resources in a consistent way, which can help data centers achieve global optimal scheduling and provide computing power with higher quality. To understand this new scheduling paradigm, in this paper, we first present an in-depth study of Alibaba's unified scheduling workloads. Our study focuses on the characterization of resource utilization, the application running performance, and scheduling scalability. We observe that although computing resources are significantly over-committed under unified scheduling, the resource utilization in Alibaba data centers is still low. In addition, existing resource usage predictors tend to make severe overestimations. At the same time, tasks within the same application behave fairly consistently, and the running performance of tasks can be well-profiled with respect to resource contention on the corresponding physical host.

Garbage Collection (GC) is a representative automatic memory manager widely deployed in popular programming languages, such as Java, C\#, and Golang (Go). Through GC, these languages provide programmers with flexibility and safety. However, GC leads to non-trivial overhead in compute and memory resources during application runtime. GC threads compete with non-GC threads (mutators) of an application, which particularly impacts latency-critical (LC) applications and causes long tail latency. Existing GC approaches do not efficiently address the interference, as GC is triggered passively without a global insight of the application; or they employ incremental GC to reduce the interference, while the incremental progress is not dynamically tailored during GC process according to runtime characteristics, which leads to significant performance degradation upon bursty requests.

Fine-tuning pre-trained language models, such as BERT, has shown enormous success among various NLP tasks. Though simple and effective, the process of fine-tuning has been found unstable, which often leads to unexpected poor performance. To increase stability and generalizability, most existing works resort to maintaining the parameters or representations of pre-trained models during fine-tuning. Nevertheless, very little work explores mining the reliable part of pre-learned information that can help to stabilize fine-tuning. To address this challenge, we introduce a novel solution in which we fine-tune BERT with stabilized cross-layer mutual information. Our method aims to preserve the reliable behaviors of cross-layer information propagation, instead of preserving the information itself, of the pre-trained model. Therefore, our method circumvents the domain conflicts between pre-trained and target tasks. We conduct extensive experiments with popular pre-trained BERT variants on NLP datasets, demonstrating the universal effectiveness and robustness of our method.

Previous work on adversarial examples typically involves a fixed norm perturbation budget, which fails to capture the way humans perceive perturbations. Recent work has shifted towards natural unrestricted adversarial examples (UAEs) that breaks ℓpperturbation bounds but nonetheless remain semantically plausible. Current methods use GAN or VAE to generate UAEs by perturbing latent codes. However, this leads to loss of high-level information, resulting in low-quality and unnatural UAEs. In light of this, we propose AdvDiffuser, a new method for synthesizing natural UAEs using diffusion models. It can generate UAEs from scratch or conditionally based on reference images. To generate natural UAEs, we perturb predicted images to steer their latent code towards the adversarial sample space of a particular classifier. We also propose adversarial inpainting based on class activation mapping to retain the salient regions of the image while perturbing less important areas. On CIFAR-10, CelebA and ImageNet, we demonstrate that it can defeat the most robust models on the RobustBench leaderboard with near 100% success rates. Furthermore, The synthesized UAEs are not only more natural but also stronger compared to the current state-of-the-art attacks. Specifically, compared with GA-attack, the UAEs generated with AdvDiffuser exhibit 6× smaller LPIPS perturbations, 2 ~ 3× smaller FID scores and 0.28 higher in SSIM metrics, making them perceptually stealthier. Finally, adversarial training with AdvDiffuser further improves the model robustness against attacks with unseen threat models.1

Transfer learning has been a popular learning paradigm in the deep learning era, especially in annotation-insufficient scenarios. Better ImageNet pre-trained models have been demonstrated, from the perspective of architecture, by previous research to have better transferability to downstream tasks[26]. However, in this paper, we find that during the same pre-training process, models at middle epochs, which are inadequately pre-trained, can outperform fully trained models when used as feature extractors (FE), while the fine-tuning (FT) performance still grows with the source performance. This reveals that there is not a solid positive correlation between top-1 accuracy on ImageNet and the transferring result on target data. Based on the contradictory phenomenon between FE and FT that a better feature extractor fails to be fine-tuned better accordingly, we conduct comprehensive analyses on features before the softmax layer to provide insightful explanations. Our discoveries suggest that, during pre-training, models tend to first learn spectral components corresponding to large singular values and the residual components contribute more when fine-tuning.

The rapid development of autonomous driving poses new research challenges to the on-vehicle computing system. In particular, the execution time of autonomous driving tasks highly depends on the specific driving environment. For instance, the execution time of configurable sensor fusion increases significantly as the scene becomes complex, which leads to end-to-end deadline misses from sensing to control and may cause accidents. Thus, a framework that can effectively utilize the system resources to guarantee the end-to-end deadlines of autonomous driving tasks as well as effectively prioritize the responsiveness and throughput of the control commands is crucial for autonomous driving. In this paper, we propose HCPerf, a performance-directed hierarchical coordination framework that intelligently coordinates the autonomous driving tasks with high execution time variation and complex dependencies according to the driving performance in real-time. Specifically, HCPerf mainly consists of two coordinators. The internal coordinator intelligently schedules the tasks according to the driving performance of the vehicle in order to help them meet the end-to-end deadlines while well prioritizing the responsiveness and throughput of the control commands. At the same time, the external coordinator dynamically tunes the rates of tasks according to the schedulability in order to efficiently utilize the system resource. We conduct extensive experiments on both simulation and hardware testbeds with the representative autonomous driving application. The results show that HCPerf can effectively improve the driving performance by 7.69%-45.94% in different driving scenarios.

Automated Feature Engineering (AFE) refers to automatically generate and select optimal feature sets for downstream tasks, which has achieved great success in real-world applications. Current AFE methods mainly focus on improving the effectiveness of the produced features, but ignoring the low-efficiency issue for large-scale deployment. Therefore, in this work, we propose a generic framework to improve the efficiency of AFE. Specifically, we construct the AFE pipeline based on reinforcement learning setting, where each feature is assigned an agent to perform feature transformation and selection, and the evaluation score of the produced features in downstream tasks serve as the reward to update the policy. We improve the efficiency of AFE in two perspectives. On the one hand, we develop a Feature Pre-Evaluation (FPE) Model to reduce the sample size and feature size that are two main factors on undermining the efficiency of feature evaluation. On the other hand, we devise a two-stage policy training strategy by running FPE on the pre-evaluation task as the initialization of the policy to avoid training policy from scratch. We conduct comprehensive experiments on 36 datasets in terms of both classification and regression tasks. The results show 2.9% higher performance in average and 2x higher computational efficiency comparing to state-of-the-art AFE methods.

Most of the existing moving object segmentation (MOS) methods regard MOS as an independent task, in this paper, we associate the MOS task with semantic segmentation, and propose a semantics-guided network for moving object segmentation (LiDAR-SGMOS). We first transform the range image and semantic features of the past scan into the range view of current scan based on the relative pose between scans. The residual image is obtained by calculating the normalized absolute difference between the current and transformed range images. Then, we apply a Meta-Kernel based cross scan fusion (CSF) module to adaptively fuse the range images and semantic features of current scan, the residual image and transformed features. Finally, the fused features with rich motion and semantic information are processed to obtain reliable MOS results. We also introduce a residual image augmentation method to further improve the MOS performance. Our method outperforms most LiDAR-MOS methods with only two sequential LiDAR scans as inputs on the SemanticKITTI MOS dataset.

A common strategy for improving efficiency in training deep learning entails multiplexing tasks on a single GPU. To mitigate the interference caused by multiplexing, existing approaches primarily employ kernel-level solutions to regulate GPU kernel execution, or harness hardware-level techniques to explicitly restrict GPU streaming multiprocessors and memory. Nevertheless, none of them perform satisfactorily in optimizing the completion time of tasks.

Abstract Real‐time individuals' destination prediction is of great significance for real‐time user tracking, service recommendation and other related applications. Traditional technology mainly used statistical methods based on the travel patterns mined from personal history travel data. However, it is not clear how to predict the destinations of individuals with only limited personal historical data. In this paper, taking the public transportation metro systems as example, we design a practical method called practical model with strong interpretability and predictability to predict each passenger's destination. Our main novelties are two aspects: (1) We propose to predict individuals' destination by combining personal and crowd behavior under certain context. (2) An explanatory model combining discrete choice model and neural network model is proposed to predict individuals' stochastic trip's destination, which can be applied to other transportation analysis scenarios about individuals' choice behavior such as travel mode choice or route choice. We validate our method based on extensive experiments, using smart card data collected by automatic fare collection system and weather data in Shenzhen, China. The experimental results demonstrate that our approach can achieve better performance than other baselines in terms of prediction accuracy.

Hybrid active and passive wireless sensor networks (HWSNs) gain advantages in extending the network lifetime and reducing the overall cost. Because the passive sensors without battery harvest the energy from distributed active sensor signal beam, and only a few active sensors can maintain a large-scale network. Thus, how to track the passive sensor’s location is essential for network management. Since the active sensors are sparsely deployed, cooperative localization which employs passive sensors to locate themselves together is a promising solution. In this article, we analyze the energy beam generated by the active sensors on the cooperative localization accuracy of the passive sensors. We consider the spatial–temporal cooperative localization based on the received signal strength (RSS) model with unknown Tx power information of each sensor due to the limited processing capabilities, circuit complexity, and energy constraints. We formulate the Fisher information matrix (FIM) and the corresponding Cramér–Rao lower bound (CRLB) for the static fully connected network and dynamic spatial–temporal recursive network. Accordingly, energy beamforming schemes are proposed to optimize localization accuracy and energy efficiency problems. For the optimal localization problem, we derive the closed-form solution of the optimal energy beamforming wave. For the optimal energy efficiency problem, we propose a semidefinite programming (SDP) solution to achieve optimal energy consumption with a self-calibration method, which can address the over-relax problem. Extensive simulation results indicate that our proposed beamforming schemes have high localization accuracy and lower power consumption compared with the existing power allocation-based schemes.

Data scarcity hinders the usability of data-dependent algorithms when tackling IoT intrusion detection (IID). To address this, we utilize the data-rich network intrusion detection (NID) domain to facilitate more accurate intrusion detection for IID domains. In this article, a geometric graph alignment (GGA) approach is leveraged to mask the geometric heterogeneities between domains for better intrusion knowledge transfer. Specifically, each intrusion domain is formulated as a graph where vertices and edges represent intrusion categories and category-wise inter-relationships, respectively. The overall shape is preserved via a confused discriminator incapable to identify adjacency matrices between different intrusion domain graphs. A rotation avoidance mechanism and a center point matching mechanism are used to avoid graph misalignment due to rotation and symmetry, respectively. Besides, category-wise semantic knowledge is transferred to act as vertex-level alignment. To exploit the target data, a pseudo-label (PL) election mechanism that jointly considers network prediction, geometric property, and neighborhood information is used to produce fine-grained PL assignment. Upon aligning the intrusion graphs geometrically from different granularities, the transferred intrusion knowledge can boost IID performance. Comprehensive experiments on several intrusion data sets demonstrate state-of-the-art performance of the GGA approach and validate the usefulness of GGA-constituting components.

As Internet of Things (IoT) devices become prevalent, using intrusion detection to protect IoT from malicious intrusions is of vital importance. However, the data scarcity of IoT hinders the effectiveness of traditional intrusion detection methods. To tackle this issue, in this article, we propose the adaptive bi-recommendation and self-improving network (ABRSI) based on unsupervised heterogeneous domain adaptation (HDA). The ABRSI transfers enrich intrusion knowledge from a data-rich network intrusion source domain to facilitate effective intrusion detection for data-scarce IoT target domains. The ABRSI achieves fine-grained intrusion knowledge transfer via adaptive bi-recommendation matching. Matching the bi-recommendation interests of two recommender systems (RSs) and the alignment of intrusion categories in the shared feature space form a mutual-benefit loop. Besides, the ABRSI uses a self-improving mechanism, autonomously improving the intrusion knowledge transfer from four ways. A hard pseudo label (PL) voting mechanism jointly considers RS decision and label relationship information to promote more accurate hard PL assignment. To promote diversity and target data participation during intrusion knowledge transfer, target instances failing to be assigned with a hard PL will be assigned with a probabilistic soft PL, forming a hybrid pseudo-labeling strategy. Meanwhile, the ABRSI also makes soft pseudo-labels globally diverse and individually certain. Finally, an error knowledge learning mechanism is utilized to adversarially exploit factors that causes detection ambiguity and learns through both current and previous error knowledge, preventing error knowledge forgetfulness. Holistically, these mechanisms form the ABRSI model that boosts IoT intrusion detection accuracy via HDA-assisted intrusion knowledge transfer. Comprehensive experiments on several intrusion data sets demonstrate the state-of-the-art performance of the ABRSI method, outperforming its counterparts by 9.2%, and also verify the effectiveness of ABRSI constituting components and ABRSI’s overall efficiency.

In this article, we propose a joint semantic transfer network (JSTN) toward effective intrusion detection (ID) for large-scale scarcely labeled Internet of Things (IoT) domain. As a multisource heterogeneous domain adaptation (MS-HDA) method, the JSTN integrates a knowledge-rich network intrusion (NI) domain and another small-scale IoT intrusion (II) domain as source domains and preserves intrinsic semantic properties to assist target II domain ID. The JSTN jointly transfers the following three semantics to learn a domain-invariant and discriminative feature representation. The scenario semantic endows source NI and II domains with characteristics from each other to ease the knowledge transfer process via a confused domain discriminator and categorical distribution knowledge preservation. It also reduces the source–target discrepancy to make the shared feature space domain invariant. Meanwhile, the weighted implicit semantic transfer boosts discriminability via a fine-grained knowledge preservation, which transfers the source categorical distribution to the target domain. The source–target divergence guides the importance weighting during knowledge preservation to reflect the degree of knowledge learning. Additionally, the hierarchical explicit semantic alignment performs centroid-level and representative-level alignment with the help of a geometric similarity-aware pseudo-label refiner, which exploits the value of the unlabeled target II domain and explicitly aligns feature representations from a global and local perspective in a concentrated manner. Comprehensive experiments on various tasks verify the superiority of the JSTN against state-of-the-art comparing methods, on average a 10.3% of accuracy boost is achieved. The statistical soundness of each constituting component and the computational efficiency is also verified.

Short-term Origin-Destination(OD) matrix prediction in metro systems aims to predict the number of passenger demands from one station to another during a short time period. That is crucial for dynamic traffic operations, e.g., route recommendation, metro scheduling. However, existing methods need further improvement due to that they fail to take full use of the real-time traffic information and model the complex spatiotemporal correlation of traffic flows. In this paper, aMulti-ViewPassengerFlow (MVPF) evolution trend based OD matrix prediction method is proposed. It consists of two components focusing on individual station and cross-station learning. Specifically, the individual station level part uses Gate Recurrent Unit and Extended Graph Attention Networks combined model to learn the high-level spatiotemporal-dependent representation of each station as the roles of origin and destination respectively, by considering multiple views of real-time traffic information (i.e., Inflow, destination allocation of Inflow, Outflow, origin allocations of Outflow). The cross-station part aims to learn passenger mobility pattern from each origin to destination through defining a transition matrix under spatiotemporal context. Compared with state-of-the-art solutions, MVPF increases the OD prediction performance metric of WMAPE by 2.5% on average. The experimental results demonstrate the superiority of MVPF against other competitors. The source code is available athttps://github.com/zfrInSIAT/MVPF-code.

In this paper, we study a data caching problem in the cloud environment, where multiple frequently co-utilised data items could be packed as a single item being transferred to serve a sequence of data requests dynamically with reduced cost. To this end, we propose an online algorithm with respect to a homogeneous cost model, calledPackCache, that can leverage the FP-Tree technique to mine those frequently co-utilised data items for packing whereby the incoming requests could be cost-effectively served online by exploiting the concept of anticipatory caching. We show the algorithm is$2/\alpha$competitive, reaching the lower bound of the competitive ratio for any deterministic online algorithm on the studied caching problem, and also time and space efficient to serve the requests. Finally, we evaluate the performance of the algorithm via experimental studies to show its actual cost-effectiveness and scalability.

Flow scheduling and congestion control are two important techniques to reduce flow completion time in data center networks. While existing works largely treat them independently, the interactions between flow scheduling and congestion control are in general overlooked which leads to sub-optimal solutions, especially given that the link capacity is increasing faster than the switch port buffer size. In this paper, we presentFlash, a simple yet effective scheme that integrates scheduling and congestion control. Specifically,Flashputs forward a congestion-aware scheduling scheme to determine the priority of flows based on the latest network congestion extent and the flow’s bytes sent. Besides,Flashproposes a priority-based packet dropping scheme in switch port buffers and implements a priority-aware congestion control scheme. Experiment results show thatFlashhas superior performance: (1) it has 35.8% lower tail latency than PIAS and performs similar with pFabric in a 10G network without knowing the flow size, (2) in 100G networks with shallow buffers, the information agnosticFlashhas 6.8% lower average FCT than the information-aware pFabric, (3) it outperforms pFabric by 13.5% in FCT if flow size is also known toFlash.

Cross-spectrum depth estimation aims to provide a reliable depth map under variant-illumination conditions with a pair of dual-spectrum images. It is valuable for autonomous driving applications when vehicles are equipped with two cameras of different modalities. However, images captured by different-modality cameras can be photometrically quite different, which makes cross-spectrum depth estimation a very challenging problem. Moreover, the shortage of large-scale open-source datasets also retards further research in this field. In this paper, we propose an unsupervised visible light(VIS)-image-guided cross-spectrum (i.e., thermal and visible-light, TIR-VIS in short) depth-estimation framework. The input of the framework consists of a cross-spectrum stereo pair (one VIS image and one thermal image). First, we train a depth-estimation base network using VIS-image stereo pairs. To adapt the trained depth-estimation network to the cross-spectrum images, we propose a multi-scale feature-transfer network to transfer features from the TIR domain to the VIS domain at the feature level. Furthermore, we introduce a mechanism of cross-spectrum depth cycle-consistency to improve the depth estimation result of dual-spectrum image pairs. Meanwhile, we release to society a large cross-spectrum dataset with visible-light and thermal stereo images captured in different scenes. The experiment result shows that our method achieves better depth-estimation results than the compared existing methods. Our code and dataset are available onhttps://github.com/whitecrow1027/CrossSP_Depth.

In this paper, we studied a data caching problem in edge-based CDNs to facilitate the content delivery to serve a sequence of requests, off-line and online, with minimum costs as a goal based on a semi-homo cost model. To this end, we first designed an O(mn \log(mn)) time and space optimal proactive off-line algorithm,called pro-caching, by reducing the problem to a simple shortest path problem in a directed weighted network graph, and then extended the idea of anticipatory caching to develop an 2-competitive reactive online algorithm, called re-caching, for this problem and showed its tightness by proving that no deterministic online algorithm can do better than 2-o(1) in its worst case. Finally, to combine the advantages of both algorithms, we also presented a hybrid algorithm, called hy-caching, to fully utilize the power and benefits of edge-based CDNs while reducing their service costs. Our results improve the previous results not only in the cost model being used but also in the time complexity, competitive ratio, and the quality of the solutions. We provably achieve these results with our deep insights into the problem and the careful analysis, together with an empirical evaluation.

With the advancement of dynamic wireless charging for Electric Vehicles (EVs), Mobile Energy Disseminator (MED), which can charge an EV in motion, becomes available. However, existing wireless charging scheduling methods for wireless sensors, which are the most related works to MED deployment, are not directly applicable for city-scale EV-to-EV dynamic wireless charging. We presentMobiCharger:aMobile wirelessChargerguidance system that determines the number of serving MEDs, and their optimal routes. We studied a metropolitan-scale vehicle mobility dataset, and found: most vehicles have routines, and the number of driving EVs changes over time, which means MED deployment should adaptively change as well. We combine EVs' current trajectories and routines to estimate EV density and the cruising graph for MED coverage. Then, we develop an offline MED deployment method that utilizes multi-objective optimization to determine the number of serving MEDs and the driving route of each MED, and an online method that utilizes Reinforcement Learning to adjust the MED deployment when the real-time vehicle traffic changes. Our trace-driven experiments show that compared with previous methods,MobiChargerincreases the medium State-of-Charge of all EVs by 50% during all time slots, and the number of charges of EVs by almost 100%.

This article presents a model-free λ -policy iteration ( λ -PI) for the discrete-time linear quadratic regulation (LQR) problem. To solve the algebraic Riccati equation arising from solving the LQR in an iterative manner, we define two novel matrix operators, named the weighted Bellman operator and the composite Bellman operator. Then, the λ -PI algorithm is first designed as a recursion with the weighted Bellman operator, and its equivalent formulation as a fixed-point iteration with the composite Bellman operator is shown. The contraction and monotonic properties of the composite Bellman operator guarantee the convergence of the λ -PI algorithm. In contrast to the PI algorithm, the λ -PI does not require an admissible initial policy, and the convergence rate outperforms the value iteration (VI) algorithm. Model-free extension of the λ -PI algorithm is developed using the off-policy reinforcement learning technique. It is also shown that the off-policy variants of the λ -PI algorithm are robust against the probing noise. Finally, simulation examples are conducted to validate the efficacy of the λ -PI algorithm.

Traffic classification is essential for cybersecurity maintenance and network management, and has been widely used in QoS (Quality of Service) guarantees, intrusion detection, and other tasks. Recently, with the emergence of SSL/TLS encryption protocols in the modern Internet environment, the traditional payload-based classification methods are no longer effective. Some researchers have used machine learning methods to model the flow features of encrypted traffics (e.g. message type, length sequence, statistical features, etc.), and achieved good results in some cases. However, these high-level hand-designed features cannot be used for more fine-grained operations and may lead to the loss of important information, thus affecting the classification accuracy. To overcome this limitation, in this paper, we designed a novel multimodal deep learning framework for encrypted traffic classification called PEAN. PEAN uses the raw bytes and length sequence as the input, and uses the self-attention mechanism to learn the deep relationship among network packets in a biflow. Furthermore, unsupervised pre-training was introduced to enhance PEAN’s ability to characterize network packets. Experiments on a real trace set captured in a large data center demonstrate the effectiveness of PEAN, which achieves better results than the state-of-the-art methods.

The rapid development of autonomous driving poses new research challenges for on-vehicle computing system. The execution time of autonomous driving tasks heavily depends on the driving environment. As the scene becomes complex, task execution time increases significantly, leading to end-to-end deadline misses and potential accidents. Hence, a framework that can effectively schedule tasks according to the driving environment in order to guarantee end-to-end deadlines is critical for autonomous driving. In this article, we propose AutoRS, an environment-dependent real-time scheduling framework for end-to-end autonomous driving. AutoRS consists of two nested control loops. The inner control loop schedules tasks based on the driving environment to help them meet end-to-end deadlines while prioritizing the responsiveness and throughput of control commands. The outer control loop tunes task rates based on schedulability to efficiently utilize system resources with an RL-based design. We conduct extensive experiments on both simulation and hardware testbeds using representative autonomous driving applications. The results demonstrate that AutoRS effectively improves the driving performance by$7.95\%-56.9\%$in different driving environments. AutoRS can significantly enhance the safety and reliability of autonomous driving systems by providing timely control commands in complex and dynamic driving environments while guaranteeing task deadlines.

With the fast growth of mobile edge computing (MEC), the deep neural network (DNN) has gained more opportunities in application to various mobile services. Given the tremendous number of learning parameters and large model size, the DNN model is often trained in cloud center and then dispatched to end devices for inference via edge network. Therefore, maximizing the cost-efficiency of learned model dispatch in the edge network would be a critical problem for the model serving in various application contexts. To reach this goal, in this article we focus mainly on reducing the total model dispatch cost in the edge network while maintaining the efficiency of the model inference. We first study this problem in its off-line form as a baseline where a sequence of$n$requests can be pre-defined in advance and exploit dynamic programming techniques to obtain a fast optimal algorithm in time complexity of$O(m^{2}n)$under a semi-homogeneous cost model in a$m$-sized network. Then, we design and implement a 2.5-competitive algorithm for its online case with a provable lower bound of 2 for any deterministic online algorithm. We verify our results through careful algorithmic analysis and validate their actual performance via a trace-based study based on a public open international mobile network dataset.

Serverless computing is growing in popularity by virtue of its lightweight and simplicity of management. It achieves these merits by reducing the granularity of the computing unit to the function level. Specifically, serverless allows users to focus squarely on the function itself while leaving other cumbersome management and scheduling issues to the platform provider, who is responsible for striking a balance between high-performance scheduling and low resource cost. In this article, we conduct a comprehensive survey of serverless computing with a particular focus on its infrastructure characteristics. Whereby some existing challenges are identified, and the associated cutting-edge solutions are analyzed. With these results, we further investigate some typical open-source frameworks and study how they address the identified challenges. Given the great advantages of serverless computing, it is expected that its deployment would dominate future cloud platforms. As such, we also envision some promising research opportunities that need to be further explored in the future. We hope that our work in this article can inspire those researchers and practitioners who are engaged in related fields to appreciate serverless computing, thereby setting foot in this promising area and making great contributions to its development.

Central to active learning (AL) is what data should be selected for annotation. Existing works attempt to select highly uncertain or informative data for annotation. Nevertheless, it remains unclear how selected data impacts the test performance of the task model used in AL. In this work, we explore such an impact by theoretically proving that selecting unlabeled data of higher gradient norm leads to a lower upper-bound of test loss, resulting in a better test performance. However, due to the lack of label information, directly computing gradient norm for unlabeled data is infeasible. To address this challenge, we propose two schemes, namely expected-gradnorm and entropy-gradnorm. The former computes the gradient norm by constructing an expected empirical loss while the latter constructs an unsupervised loss with entropy. Furthermore, we integrate the two schemes in a universal AL framework. We evaluate our method on classical image classification and semantic segmentation tasks. To demonstrate its competency in domain applications and its robustness to noise, we also validate our method on a cellular imaging analysis task, namely cryo-Electron Tomography subtomogram classification. Results demonstrate that our method achieves superior performance against the state of the art. We refer readers to https://arxiv.org/pdf/2112.05683.pdf for the full version of this paper which includes the appendix and source code link.

The cold-start problem, faced when providing recommendations to newly joined users with no historical interaction record existing in the platform, is one of the most critical problems that negatively impact the performance of a recommendation system. Fortunately, cross-domain recommendation~(CDR) is a promising approach for solving this problem, which can exploit the knowledge of these users from source domains to provide recommendations in the target domain. However, this method requires that the central server has the interaction behaviour data in both domains of all the users, which prevents users from participating due to privacy issues.

When deploying microservices in production clusters, it is critical to automatically scale containers to improve cluster utilization and ensure service level agreements (SLA). Although reactive scaling approaches work well for monolithic architectures, they are not necessarily suitable for microservice frameworks due to the long delay caused by complex microservice call chains. In contrast, existing proactive approaches leverage end-to-end performance prediction for scaling, but cannot effectively handle microservice multiplexing and dynamic microservice dependencies.

Federated learning (FL) allows multiple clients to collectively train a high-performance global model without sharing their private data. However, the key challenge in federated learning is that the clients have significant statistical heterogeneity among their local data distributions, which would cause inconsistent optimized local models on the clientside. To address this fundamental dilemma, we propose a novel federated learning algorithm with local drift decoupling and correction (FedDC). Our FedDC only introduces lightweight modifications in the local training phase, in which each client utilizes an auxiliary local drift variable to track the gap between the local model parameter and the global model parameters. The key idea of FedDC is to utilize this learned local drift variable to bridge the gap, i.e., conducting consistency in parameter-level. The experiment results and analysis demonstrate that FedDC yields expediting convergence and better performance on various image classification tasks, robust in partial participation settings, non-iid data, and heterogeneous clients.

Position Embedding (PE) is essential for transformer to capture the sequence ordering of input tokens.Despite its general effectiveness verified in Natural Language Processing (NLP) and Computer Vision (CV), its application in cross-modal tasks remains unexplored and suffers from two challenges: 1) the input text tokens and image patches are not aligned; 2) the encoding space of each modality is different, making it unavailable for feature comparison.In this paper, we propose a unified position embedding method for these problems, called AnChor-basEd Relative Position Embedding (ACE-RPE), in which we first introduce an anchor locating mechanism to bridge the semantic gap and locate anchors from different modalities.Then we conduct the distance calculation of each text token and image patch by computing their shortest paths from the located anchors.Last, we embed the anchor-based distance to guide the computation of crossattention.In this way, it calculates cross-modal relative position embeddings for cross-modal transformer.Benefiting from ACE-RPE, our method obtains new SOTA results on a wide range of benchmarks, such as Image-Text Retrieval on MS-COCO and Flickr30K, Visual Entailment on SNLI-VE, Visual Reasoning on NLVR2 and Weakly-supervised Visual Grounding on RefCOCO+.

One of the most popular location privacy-preserving mechanisms applied in location-based services (LBS) is location obfuscation, where mobile users are allowed to report obfuscated locations instead of their real locations to services. Many existing obfuscation approaches consider mobile users that can move freely over a region. However, this is inadequate for protecting the location privacy of vehicles, as their mobility is restricted by external factors, such as road networks and traffic flows. This auxiliary information about external factors helps an attacker to shrink the search range of vehicles' locations, increasing the risk of location exposure.

Co-location of latency-critical services with best-effort batch jobs is commonly adopted in production systems to increase resource utilization. Although memory and CPU isolation have been extensively studied, we find Simultaneous Multi-Threading (SMT) technology imposes non-trivial interference on memory access which jeopardizes efficient co-location and performance assurance of latency-critical services. However, there is not an existing metric to quantitatively measure and lacks a deterministic approach to tackle SMT interference on memory access.

With the development of vehicular networks, the vehicle-to-everything (V2X) communication aided emergency warning is envisioned to improve the safety of driving service in intelligent transportation systems (ITS). By considering the delay sensitivity of different vehicles receiving warning information, this paper investigates the V2X communication aided emergency message dissemination. Specifically, according to the distance between the vehicle and the emergency point, we divide the vehicles in the coverage of the roadside unit (RSU) into two groups, namely, the primary priority group and the secondary priority group. Then, we formulate a joint optimization problem for content partition, user grouping and channel allocation to improve the resource utilization and the efficiency of emergency message delivery. The objective is to ensure that all vehicles in the primary priority group can reliably receive the warning messages within a fixed deadline, and meanwhile, the RSU can send as many warning messages as possible to the vehicles in the secondary priority group. Despite the nature of mixed integer and non-linear programming problem, we propose a layered approach to solve the problem. Finally, we conduct simulations to validate the efficiency and effectiveness of the proposed algorithm, compared to some benchmark algorithms.

To enable self-driving without a human driver, an autonomous vehicle needs to perceive its surrounding obstacles using onboard sensors, of which the perception accuracy might be limited by their own sensing range. An effective way to improve vehicles’ perception accuracy is to let nearby vehicles exchange their sensor data so that vehicles can detect obstacles beyond their own sensing ranges, called cooperative perception. The shared sensor data, however, might disclose the sensitive information of vehicles’ passengers, raising privacy and safety concerns (e.g. stalking or sensitive location leakage).In this paper, we propose a new data-sharing policy for the cooperative perception of autonomous vehicles, of which the objective is to minimize vehicles’ information disclosure without compromising their perception accuracy. Considering vehicles usually have different desires for data-sharing under different traffic environments, our policy provides vehicles autonomy to determine what types of sensor data to share based on their own needs. Moreover, given the dynamics of vehicles’ data-sharing decisions, the policy can be adjusted to incentivize vehicles’ decisions to converge to the desired decision field, such that a healthy cooperation environment can be maintained in a long term. To achieve such objectives, we analyze the dynamics of vehicles’ data-sharing decisions by resorting to the game theory model, and optimize the data-sharing ratio in the policy based on the analytic results. Finally, we carry out an extensive trace-driven simulation to test the performance of the proposed data-sharing policy. The experimental results demonstrate that our policy can help incentivize vehicles’ data-sharing decisions to the desired decision fields efficiently and effectively.

Place recognition or loop closure detection is one of the core components in a full SLAM system. In this paper, aiming at strengthening the relevancy of local neighboring points and the contextual dependency among global points simultaneously, we investigate the exploitation of transformer-based network for feature extraction, and propose a Hierarchical Transformer for Place Recognition (HiTPR). The HiTPR consists of four major parts: point cell generation, short-range transformer (SRT), long-range transformer (LRT) and global descriptor aggregation. Specifically, the point cloud is initially divided into a sequence of small cells by down-sampling and nearest neighbors searching. In the SRT, we extract the local feature for each point cell. While in the LRT, we build the global dependency among all of the point cells in the whole point cloud. Experiments on several standard benchmarks demonstrate the superiority of the HiTPR in terms of average recall rate, achieving 93.71 % at top 1 % and 86.63 % at top 1 on the Oxford RobotCar dataset for example.

In this paper, we propose an efficient frontier detector method based on adaptive Rapidly-exploring Random Tree (RRT) for autonomous robot exploration. Robots can achieve real-time incremental frontier detection when they are exploring unknown environments. First, our detector adaptively adjusts the sampling space of RRT by sensing the surrounding environment structure. The adaptive sampling space can greatly improve the successful sampling rate of RRT (the ratio of the number of samples successfully added to the RRT tree to the number of sampling attempts) according to the environment structure and control the expansion bias of the RRT. Second, by generating non-uniform distributed samples, our method also solves the over-sampling problem of RRT in the sliding windows, where uniform random sampling causes over-sampling in the overlap area between two adjacent sliding windows. In this way, our detector is more inclined to sample in the latest explored area, which improves the efficiency of frontier detection and achieves incremental detection. We validated our method in three simulated benchmark scenarios. The experimental comparison shows that we reduce the frontier detection runtime by about 40% compared with the SOTA method, DSV Planner.

Grounding a command to the visual environment is an essential ingredient for interactions between autonomous vehicles and humans. In this work, we study the problem of language grounding for autonomous vehicles, which aims to localize a region in a visual scene according to a natural language command from a passenger. Prior work only employs the top layer representations of a vision-and-language pretrained model to predict the region referred to by the command. However, such a method omits the useful features encoded in other layers, and thus results in inadequate understanding of the input scene and command. To tackle this limitation, we present the first layer fusion approach for this task. Since different visual regions may require distinct types of features to disambiguate them from each other, we further propose the region-specific dynamic (RSD) layer attention to adaptively fuse the multimodal information across layers for each region. Extensive experiments on the Talk2Car benchmark demonstrate that our approach helps predict more accurate regions and outperforms state-of-the-art methods.

In this paper, we propose a learning-based moving-object tracking method utilizing the newly developed LiDAR sensor, Frequency Modulated Continuous Wave (FMCW) LiDAR. Compared with most existing commercial LiDAR sensors, FMCW LiDAR can provide additional Doppler velocity information to each 3D point of the point clouds. Benefiting from this, we can generate instance labels as ground truth in a semi-automatic manner. Given the labels, we propose a contrastive learning framework, which pulls together the features from the same instance in embedding space and pushes apart the features from different instances, to improve the tracking quality. Extensive experiments are conducted on the recorded driving data, and the results show that our method outperforms the baseline methods by a large margin.

Federated learning (FL) enables multiple devices to collaboratively train a shared model while preserving data privacy. However, despite its emerging applications in many areas, real-world deployment of on-device FL is challenging due to wildly diverse training capability and data distribution across heterogeneous edge devices, which highly impact both model performance and training efficiency. This paper proposes Harmony, a high-performance FL framework with heterogeneity-aware hierarchical management of training devices and training data. Unlike previous work that mainly focuses on heterogeneity in either training capability or data distribution, Harmony adopts a hierarchical structure to jointly handle both heterogeneities in a unified manner. Specifically, the two core components of Harmony are a global coordinator hosted by the central server and a local coordinator deployed on each participating device. Without accessing the raw data, the global coordinator first selects the participants, and then further reorganizes their training samples based on the accurate estimation of the runtime training capability and data distribution of each device. The local coordinator keeps monitoring the local training status and conducts efficient training with guidance from the global coordinator. We conduct extensive experiments to evaluate Harmony using both hardware and simulation testbeds on representative datasets. The experimental results show that Harmony improves the accuracy performance by 1.67% - 27.62%. In addition, Harmony effectively accelerates the training process up to $3.29\times$ and $1.84\times$ on average, and saves energy up to 88.41% and 28.04% on average.

For resource utilization efficiency, latency critical (LC) services are commonly co-located with best-effort batch jobs in datacenter servers. Many LC services, such as Cassandra and HBase, run in Java Virtual Machine (JVM). We find that LC services often experience heavy-tailed latency due to performance interference of the concurrent garbage collection (GC) as well as multi-tenancy. The root cause is a semantic gap of resource allocation between JVM and the underlying Linux OS in multi-tenant systems. That is, the OS is unaware of the characteristics of different kinds of threads in JVM (i.e., GC threads and LC worker threads), which may lead to GC threads competing for CPUs; JVM is unaware of the resource utilization in the OS, which may trigger CPU-intensive GC operations when CPUs are busy. Furthermore, we find that co-located batch jobs can interfere with LC services due to Simultaneous Multi-Threading (SMT).

Adversarial attacks can deceive neural networks by adding tiny perturbations to their input data. Ensemble defenses, which are trained to minimize attack transferability among sub-models, offer a promising research direction to improve robustness against such attacks while maintaining a high accuracy on natural inputs. We discover, however, that recent state-of-the-art (SOTA) adversarial attack strategies cannot reliably evaluate ensemble defenses, sizeably overestimating their robustness. This paper identifies the two factors that contribute to this behavior. First, these defenses form ensembles that are notably difficult for existing gradient-based method to attack, due to gradient obfuscation. Second, ensemble defenses diversify sub-model gradients, presenting a challenge to defeat all sub-models simultaneously, simply summing their contributions may counteract the overall attack objective; yet, we observe that ensemble may still be fooled despite most sub-models being correct. We therefore introduce MORA, a model-reweighing attack to steer adversarial example synthesis by reweighing the importance of sub-model gradients. MORA finds that recent ensemble defenses all exhibit varying degrees of overestimated robustness. Comparing it against recent SOTA white-box attacks, it can converge orders of magnitude faster while achieving higher attack success rates across all ensemble models examined with three different ensemble modes (i.e, ensembling by either softmax, voting or logits). In particular, most ensemble defenses exhibit near or exactly $0\%$ robustness against MORA with $\ell^\infty$ perturbation within $0.02$ on CIFAR-10, and $0.01$ on CIFAR-100. We make MORA open source with reproducible results and pre-trained models; and provide a leaderboard of ensemble defenses under various attack strategies.

Summary With the fast growing quantity of data generated by smart devices and the exponential surge of processing demand in the Internet of Things (IoT) era, the resource‐rich cloud centers have been utilized to tackle these challenges. To relieve the burden on cloud centers, edge‐cloud computation offloading becomes a promising solution since shortening the proximity between the data source and the computation by offloading computation tasks from the cloud to edge devices can improve performance and quality of service. Several optimization models of edge‐cloud computation offloading have been proposed that take computation costs and heterogeneous communication costs into account. However, several important factors are not jointly considered, such as heterogeneities of tasks, load balancing among nodes and the profit yielded by computation tasks, which lead to the profit and cost‐oriented computation offloading optimization modelPECCOproposed in this article. Considering that the model is hard in nature and the optimization objective is not differentiable, we propose an improved Moth‐flame optimizerPECCO‐MFIwhich addresses some deficiencies of the original Moth‐flame optimizer and integrate it under the edge‐cloud environment. Comprehensive experiments are conducted to verify the superior performance of the proposed method when optimizing the proposed task offloading model under the edge‐cloud environment.

Abstract As one of the most useful online processing techniques, the theta‐join operation has been utilized by many applications to fully excavate the relationships between data streams in various scenarios. As such, constant research efforts have been put to optimize its performance in the distributed environment, which is typically characterized by reducing the number of Cartesian products as much as possible. In this article, we design and implement a novel fast theta‐join algorithm, calledPrefap, by developing two distinct techniques—prefilteringandamalgamated partitioning—based on the state‐of‐the‐art FastThetaJoin algorithm to optimize the efficiency of the theta‐join operation. Firstly, we develop a prefiltering strategy before data streams are partitioned to reduce the amount of data to be involved and benefit a more fine‐grained partitioning. Secondly, to avoid the data streams being partitioned in a coarse‐grained isolated manner and improve the quality of the partition‐level filtering, we introduce an amalgamated partitioning mechanism that can amalgamate the partitioning boundaries of two data streams to assist a fine‐grained partitioning. With the integration of these two techniques into the existing FastThetaJoin algorithm, we design and implement a new framework to achieve a decreased number of Cartesian products and a higher theta‐join efficiency. By comparing with existing algorithms, FastThetaJoin in particular, we evaluate the performance ofPrefapon both synthetic and real data streams from two‐way to multiway theta‐join to demonstrate its superiority.

Abstract Semantic segmentation is a hot topic in computer vision and various deep learning networks are designed to achieve higher accuracy on that by fully exploring the capability of neural networks. This paper aims to address the issue and proposes the substructures with novelty for popular networks. Meanwhile, we present a cross‐channel structure, which simultaneously reduces parameter while the kernel size becomes larger. After that, to overcome the weakness of insufficient dataset which refers to satellite image data, we propose a feature attention mechanism with generative adversarial network to enhance the images' features. We show the recognition result on the satellite image dataset with a large picture. This paper evaluates substructures on the PASCAL VOC2012 dataset and improves the mIOU from 74.68% to 88.15%.

Previous methods for passenger demand inference are unable to capture the effect of all possible random factors (e.g., accident and weather), hence resulting in insufficient accuracy. Moreover, due to the lack of charging optimization, existing taxicab dispatching methods cannot be applied to electric taxicabs directly. We propose CD-Guide, which provides Charging and Dispatching Guide for electric taxicabs based on customized selection and training of historical passenger demand data, multiobjective optimization, and reinforcement learning (RL). By analyzing a large-scale electric taxicab data set, we found that: 1) the histogram of passengers’ origin buildings is effective in illustrating the suitability of historical data for learning; 2) passenger demands in different regions vary a lot due to various random factors; and 3) charging time must be considered in dispatching electric taxicabs. We first develop a passenger demand inference model based on customized selection and training of suitable historical passenger demand data. Then, we develop two taxicab guidance methods that utilize multiobjective optimization and RL, respectively, to maximize the taxicab’s likelihood of finding passengers, maximally prevent the taxicab from missing passengers due to charging, and, meanwhile, maintain the continuous service of the taxicab. Extensive experiments on real-world data sets demonstrate that compared with the state of the art, CD-Guide increases the total number of served passengers by 100%, and the minimum State-of-Charge of all taxicabs by 75% during all time slots.

In metropolitan areas with heavy transit demands, electric vehicles (EVs) are expected to be continuously driving without recharging downtime. Wireless power transfer (WPT) provides a promising solution for in-motion EV charging. Nevertheless, previous works are not directly applicable for the deployment of in-motion wireless chargers due to their different charging characteristics. The challenge of deploying in-motion wireless chargers to support the continuous driving of EVs in a metropolitan road network with the minimum cost remains unsolved. We proposeCatChargerto tackle this challenge. By analyzing a metropolitan-scale data set, we found that traffic attributes like vehicle passing speed, daily visit frequency at intersections (i.e., landmarks), and their variances are diverse, and these attributes are critical to in-motion wireless charging performance. Driven by these observations, we first group landmarks with similar attribute values using the entropy minimization clustering method, and select candidate landmarks from the groups with suitable attribute values. Then, we use the kernel density estimator (KDE) to deduce the expected vehicle residual energy at each candidate landmark and consider EV drivers’ routing choice behavior in charger deployment. Finally, we determine the deployment locations by formulating and solving a multiobjective optimization problem, which maximizes vehicle traffic flow at charger deployment positions while guaranteeing the continuous driving of EVs at each landmark. Trace-driven experiments demonstrate thatCatChargerincreases the ratio of driving EVs at the end of a day by 12.5% under the same deployment cost.

Abstract As a key technology in the 5G era, mobile edge computing (MEC) has developed rapidly in recent years. MEC aims to reduce the service delay of mobile users, while alleviating the processing pressure on the core network. MEC can be regarded as an extension of cloud computing on the user side, which can deploy edge servers and bring computing resources closer to mobile users, and provide more efficient interactions. However, due to the user's dynamic mobility, the distance between the user and the edge server will change dynamically, which may cause fluctuations in Quality of Service. Therefore, when a mobile user moves in the MEC environment, certain approaches are needed to schedule services deployed on the edge server to ensure the user experience. In this article, we model service scheduling in MEC scenarios and propose a delay‐aware and mobility‐aware service management approach based on concise probabilistic methods. This approach has low computational complexity and can effectively reduce service delay and migration costs. Furthermore, we conduct experiments by utilizing multiple realistic datasets and use iFogSim to evaluate the performance of the algorithm. The results show that our proposed approach can optimize the performance on service delay, with 8%–20% improvement and reduce the migration cost by more than 75% compared with baselines during the rush hours.

We present an extension of the bankers algorithm to resolve deadlock for programs whose resource-request graph can be modeled as a recursion tree for parallel execution. Our algorithm implements the bankers logic, with the key difference being that some properties of the tree are fully exploited to improve the resource utilization and safety check in deadlock avoidance. For an n-node tree modeled program making requests to m types of resources, our recursion-tree based algorithm can obtain a time complexity of O(mn loglogn) on average in safety check while reducing the conservativeness in resource utilization. We reap these benefits by proposing a concept of the resource critical tree and leverage it to localize the maximum claim associated with each node in the tree. To tackle the case when the tree model is not statically known, we relax the definition of a local maximum claim by sacrificing some resource utilization. With this trade-off, the algorithm can resolve the deadlock and achieve more efficient safety checks within time of O(m loglogn). Our empirical studies on a two-dimensional integration problem on sparse grids show that the proposed algorithms can reduce resource utilization conservativeness and improve avoidance performance by minimizing the number of safety checks.

Mobile edge computing is emerging to provide cloud-computing capabilities to mobile users, so that they can offload computation intensive tasks to close proximity for execution. However, most existing works imply that a transmission-finished task still occupies the channel until all users on the same channel finish the transmission, leading to severe channel resource waste. To solve this problem, we propose an efficient computation offloading mechanism which releases the channel resources of transmission-finished tasks for transmission-unfinished tasks, and aims to minimize the response time and energy consumption for each user. Specifically, we formulate the computation offloading problem as a game, analyze its structural properties and show how it possesses a Nash equilibrium and admits the finite improvement property, in the cases of elastic cloud and non-elastic cloud respectively. We then propose aDistributedMulti-channelComputationOffloading (DMCO) algorithm, which can converge to a Nash equilibrium, and find the upper bound of the convergence time. We further evaluate the performance of DMCO using the price of anarchy. Numerical results show that DMCO scales well with the number of users, and outperforms existing works, for example, benefits 13.3 percent more users and reduces cost by 23.7 percent than CO, one of the best existing works.

Reducing tail latency becomes increasingly important to improve the user-perceived service experience. User-facing latency-sensitive cloud applications typically contain multiple interactive tiers (e.g., Web, App, Database) running in different virtual machines (VMs) with complex interaction patterns. However, such interactions between VMs in different tiers are often neglected in previous VM consolidation methods, resulting in poor application performance. In this article, we study the consolidation of multi-tier interactive workloads from a new perspective of user-perceived tail latency. We propose a novel profiling-based consolidation methodology to satisfy tail latency requirements while reducing the number of used physical machines. To achieve such a goal, we first perform large-scale profiling experiments under various consolidation settings in a KVM virtualized private cluster to establish the empirical performance values. We consider two key factors that affect the tail latency of multi-tier workloads:interferencewith co-located VMs andinteractionbetween tiers. We model the consolidation of multi-tier workloads as an optimization problem with different objectives and constraints, and derive the consolidation schedule. We implement and evaluate the proposed models, as well as comparing with other methods (i.e.,withoutprofiling orwithoutconsidering interaction influence). Extensive experimental results show that the proposed method is able to reduce up to5Xtail latency, compared with the methodwithoutprofiling and up to1.3Xtail latency, compared with the methodwithoutconsidering the interaction influence between different tiers.

In this article, we consider an iterative adaptive dynamic programming (ADP) algorithm within the Hamiltonian-driven framework to solve the Hamilton-Jacobi-Bellman (HJB) equation for the infinite-horizon optimal control problem in continuous time for nonlinear systems. First, a novel function, "min-Hamiltonian," is defined to capture the fundamental properties of the classical Hamiltonian. It is shown that both the HJB equation and the policy iteration (PI) algorithm can be formulated in terms of the min-Hamiltonian within the Hamiltonian-driven framework. Moreover, we develop an iterative ADP algorithm that takes into consideration the approximation errors during the policy evaluation step. We then derive a sufficient condition on the iterative value gradient to guarantee closed-loop stability of the equilibrium point as well as convergence to the optimal value. A model-free extension based on an off-policy reinforcement learning (RL) technique is also provided. Finally, numerical results illustrate the efficacy of the proposed framework.

This article considers the robust optimal control problem for a class of nonlinear systems in the presence of unmodeled dynamics. An adaptive optimal controller is designed using the online actor–critic learning and is robustified against unmodeled dynamics. To deal with unmodeled dynamics, an auxiliary signal with the system state as its input signal is designed to capture the input-to-state stability. In addition to the critic network for value function approximation, a novel robustifying term is developed and introduced into the actor network to ensure robustness during the learning process. It is shown that both the actor and the critic weights learning converge to their optimal values while guaranteeing the boundedness of all the signals in the closed loop. Simulation examples are conducted to verify the efficacy of the presented scheme.

This article considers the distributed adaptive neuro-fuzzy output feedback control protocol design to solve the output synchronization problem for heterogeneous multiagent systems with nonlinear strict-feedback agent dynamics. The output constraints and actuator saturation are considered simultaneously. First, a distributed high-gain observer is employed to estimate the unmeasured agent state and relax the requirement of the Lipschitz continuity of nonlinear follower dynamics. Second, an asymmetric barrier Lyapunov function with time-varying constraint is presented to deal with both the transient and the steady-state constraints on the output synchronization error. To avoid the “explosion of complexity,” the dynamic surface control technique is employed to filter the virtual control signal for each follower. To deal with the actuator saturation, a distributed auxiliary dynamical system is designed for each follower. The fuzzy logic system is employed to compensate for the uncertain follower dynamics with guaranteed semiglobal uniformly ultimately boundedness of all closed-loop signals. Finally, a simulation example is conducted to verify the efficacy of the presented adaptive neuro-fuzzy controller design.

In recent years, various deep learning architectures have been proposed to solve complex challenges (e.g. spatial dependency, temporal dependency) in traffic domain, which have achieved satisfactory performance. These architectures are composed of multiple deep learning techniques in order to tackle various challenges in traffic tasks. Traditionally, convolution neural networks (CNNs) are utilized to model spatial dependency by decomposing the traffic network as grids. However, many traffic networks are graph-structured in nature. In order to utilize such spatial information fully, it’s more appropriate to formulate traffic networks as graphs mathematically. Recently, various novel deep learning techniques have been developed to process graph data, called graph neural networks (GNNs). More and more works combine GNNs with other deep learning techniques to construct an architecture dealing with various challenges in a complex traffic task, where GNNs are responsible for extracting spatial correlations in traffic network. These graph-based architectures have achieved state-of-the-art performance. To provide a comprehensive and clear picture of such emerging trend, this survey carefully examines various graph-based deep learning architectures in many traffic applications. We first give guidelines to formulate a traffic problem based on graph and construct graphs from various kinds of traffic datasets. Then we decompose these graph-based architectures to discuss their shared deep learning techniques, clarifying the utilization of each technique in traffic tasks. What’s more, we summarize some common traffic challenges and the corresponding graph-based deep learning solutions to each challenge. Finally, we provide benchmark datasets, open source codes and future research directions in this rapidly growing field.

Traditionally, road detection approaches mostly capitalize on RGB images, 3D LiDAR point cloud or their fusion. However, RGB camera is sensitive to light conditions, while LiDAR point cloud is sparse compared with dense image pixels. In this work, a new hybrid image dataset is provided for the task of road detection based on cameras. In this dataset, the hybrid images are acquired by an optically aligned hybrid imaging device, consisting of a far-infrared (FIR) imager and an RGB camera to output pixel-wise registration of thermal and RGB frames. Then we investigate on three methods based on fully convolutional neural network (F-CNN) to demonstrate the advantages by fusing RGB-FIR images in road detection. First, a middle-fusion based model is built, where the output feature maps of encoder branches from RGB and FIR images are directly concatenated into a single-fusion branch as the decoder. Next, the originally discarded layers after fusion operation for both RGB and FIR branches are recovered as the mimic branches to imitate the distributions of the fusion outputs, which constitutes an extended cross model (ECM). Moreover, the outputs of mimic branches at different scales are also used to imitate the corresponding outputs in the fusion branch, called a hierarchical cross model (HCM). The experimental results demonstrate the effectiveness and efficiency of our fusion strategies.

Understanding and predicting each individual’s real-time travel destination given the origin information in urban public transportation systems is crucial for personalized traveler recommendation, targeted demand management, dynamic traffic operations and so on. Existing methods are often based on modeling the regular travel patterns through analyzing the long-term personal travel information. They are suitable for destination prediction of individual regular trips with regular travel patterns, but may not work well for occasional trips with strong randomness and uncertainty, especially for the individuals with a few historical travel data. In this paper, we focus on more challenging issue about destination prediction of occasional trips. We design a general Multi-View Deep Learning Framework (MDLF) based on the data-driven insight that a location where a user will destine to is not only related to the user’s own travel preference to the location, but also influenced by crowd’s travel preference and the region’s characteristics of the location under certain spatiotemporal contexts. The destination of an individual’s occasional trip can be predicted by combining all these complementary influencing factors. The novelty of MDLF is mainly reflected in two aspects. The first is the effective feature extraction from multiple and complementary views. The second is that a CNN (Recurrent Neural Network) based deep learning component for predicting each occasional trip’s destination by calculating a moving trend score for each possible destination. We evaluate the MDLF based on two real-world smart card datasets collected by AFC (Automatic Fare Collection) Systems. The experimental results demonstrate the superiority of MDLF against other competitors.

In this paper, we investigate an efficient way for identifying passengers in a metro system across two heterogeneous but complementary trajectory data sources: AFC data recording two points per trip about when and where a passenger enters or leaves the metro system, and WiFi data recording a few points passed by a passenger in the way of some of his/her trips. The identification result can help us to complete individuals’ mobility, and benefits to lots of services, e.g., individual route choice analysis, epidemic case detecting and so on. The problem is similar to calculate the similarity between two trajectories from two data sources, where a trajectory refers to a sequence of points where a passenger appeared in a metro system on observed days. However, due to the small location space in a metro network, large number of passengers with similar travel pattern, and so on, there are lots of trip overlaps or point co-occurrences between different passengers. That results in a large number of passengers mismatched by existing trajectory similarity measurement. To address the problem, this paper proposes a novel global-local correlation based trajectory similarity measurement GLTC. Specifically, GLTC first extracts all overlapping trip pairs of two trajectories by considering the spatiotemporal inclusions from global level. Then it gets the similarity by aggregating each overlapping trip pair’s local similarity, which is calculated by considering some data-driven insights helpful to uniquely identify a passenger (e.g., uneven passenger flow distribution in different cross-sections of a metro network, the number of trips in same travel pattern of a trajectory, and so on). We evaluate GLTC based on real-world data, and the experimental result shows that GLTC outperforms other baselines.

Transfer learning through fine-tuning a pre-trained neural network with an extremely large dataset, such as ImageNet, can significantly improve and accelerate training while the accuracy is frequently bottlenecked by the limited dataset size of the new target task. To solve the problem, some regularization methods, constraining the outer layer weights of the target network using the starting point as references (SPAR), have been studied. In this article, we propose a novel regularized transfer learning framework \operatorname{DELTA} , namely DE ep L earning T ransfer using Feature Map with A ttention . Instead of constraining the weights of neural network, \operatorname{DELTA} aims at preserving the outer layer outputs of the source network. Specifically, in addition to minimizing the empirical loss, \operatorname{DELTA} aligns the outer layer outputs of two networks, through constraining a subset of feature maps that are precisely selected by attention that has been learned in a supervised learning manner. We evaluate \operatorname{DELTA} with the state-of-the-art algorithms, including L^2 and \emph {L}^2\text{-}SP . The experiment results show that our method outperforms these baselines with higher accuracy for new tasks. Code has been made publicly available. 1

The emerging trend towards moving from monolithic applications to microservices has raised new performance challenges in cloud computing environments. Compared with traditional monolithic applications, the microservices are lightweight, fine-grained, and must be executed in a shorter time. Efficient scaling approaches are required to ensure microservices’ system performance under diverse workloads with strict Quality of Service (QoS) requirements and optimize resource provisioning. To solve this problem, we investigate the trade-offs between the dominant scaling techniques, including horizontal scaling, vertical scaling, and brownout in terms of execution cost and response time. We first present a prediction algorithm based on gradient recurrent units to accurately predict workloads assisting in scaling to achieve efficient scaling. Further, we propose a multi-faceted scaling approach using reinforcement learning called CoScal to learn the scaling techniques efficiently. The proposed CoScal approach takes full advantage of data-driven decisions and improves the system performance in terms of high communication cost and delay. We validate our proposed solution by implementing a containerized microservice prototype system and evaluated with two microservice applications. The extensive experiments demonstrate that CoScal reduces response time by 19%-29% and decreases the connection time of services by 16% when compared with the state-of-the-art scaling techniques for Sock Shop application. CoScal can also improve the number of successful transactions with 6%-10% for Stan’s Robot Shop application.

Cloud computing has been regarded as a successful paradigm for IT industry by providing benefits for both service providers and customers. In spite of the advantages, cloud computing also suffers from distinct challenges, and one of them is the inefficient resource provisioning for dynamic workloads. Accurate workload predictions for cloud computing can support efficient resource provisioning and avoid resource wastage. However, due to the high-dimensional and high-variable features of cloud workloads, it is difficult to predict the workloads effectively and accurately. The current dominant work for cloud workload prediction is based on regression approaches or recurrent neural networks, which fail to capture the long-term variance of workloads. To address the challenges and overcome the limitations of existing works, we proposed an e fficient supervised learning-based D eep N eural Network ( esDNN ) approach for cloud workload prediction. First, we utilize a sliding window to convert the multivariate data into a supervised learning time series that allows deep learning for processing. Then, we apply a revised Gated Recurrent Unit (GRU) to achieve accurate prediction. To show the effectiveness of esDNN, we also conduct comprehensive experiments based on realistic traces derived from Alibaba and Google cloud data centers. The experimental results demonstrate that esDNN can accurately and efficiently predict cloud workloads. Compared with the state-of-the-art baselines, esDNN can reduce the mean square errors significantly, e.g., 15%. rather than the approach using GRU only. We also apply esDNN for machines auto-scaling, which illustrates that esDNN can reduce the number of active hosts efficiently, thus the costs of service providers can be optimized.

File system metadata is the data in charge of maintaining namespace, permission semantics and location of file data blocks. Operations on the metadata can account for up to 80% of total file system operations. As such, the performance of metadata services significantly impacts the overall performance of file systems. A large-scale distributed file system (DFS) is a storage system that is composed of multiple storage devices spreading across different sites to accommodate data files, and in most cases, to provide users with location independent access interfaces. Large-scale DFSs have been widely deployed as a substrate to a plethora of computing systems, and thus their metadata management efficiency is crucial to a massive number of applications, especially with the advent of the Big Data age, which poses tremendous pressure on underlying storage systems. This paper reports the state-of-the-art research on metadata services in large-scale distributed file systems, which is conducted from three indicative perspectives that are always used to characterize DFSs: high-scalability, high-performance, and high-availability, with special focus on their respective major challenges as well as their developed mainstream technologies. Additionally, the paper also identifies and analyzes several existing problems in the research, which could be used as a reference for related studies.

Loosely-coupled and light-weight microservices running in containers are replacing monolithic applications gradually. Understanding the characteristics of microservices is critical to make good use of microservice architectures. However, there is no comprehensive study about microservice and its related systems in production environments so far. In this paper, we present a solid analysis of large-scale deployments of microservices at Alibaba clusters. Our study focuses on the characterization of microservice dependency as well as its runtime performance. We conduct an in-depth anatomy of microservice call graphs to quantify the difference between them and traditional DAGs of data-parallel jobs. In particular, we observe that microservice call graphs are heavy-tail distributed and their topology is similar to a tree and moreover, many microservices are hot-spots. We also discover that the structure of call graphs for long-term developed applications is much simpler so as to provide better performance. Our investigation on microservice runtime performance indicates most microservices are much more sensitive to CPU interference than memory interference. Moreover, we design resource management policies to efficiently tune memory resources.

Urban traffic network is a typical complex network. Traffic states data (e.g., traffic flow, traffic occupancy, traffic speed, etc.) has strong temporal and spatial correlation. To accurately predict urban traffic state, it is very important to extract the road features in the traffic network. The existing methods use separated temporal and spatial components or Spatio-temporal fusion components to predict traffic. Graph Convolution Network (GCN) is usually used to obtain the correlation between spatial nodes or Spatio-temporal nodes. However, the message aggregation method of GCN cannot assign different weights to neighbor nodes. While Graph Attention Network (GAT) can pay attention to different neighbor nodes. To better explain the existing traffic prediction models, we carried out experiments on the model framework based on GCN. We use a new proposed GAT instead of GCN, and find that the new GAT has better performance in multi-features traffic prediction tasks. We also made a theoretical analysis on the improvement of the performance and carried out experiments on four real datasets, which can provide strong support for the theoretical analysis. Our method improves the interpretability of the Graph Neural Network (GNN) model in extracting spatial features of the traffic networks.

Loosely-coupled and light-weight microservices running in containers are replacing monolithic applications gradually. Understanding the characteristics of microservices is critical to make good use of microservice architectures. However, there is no comprehensive study about microservice and its related systems in production environments so far. In this paper, we present a solid analysis of large-scale deployments of microservices at Alibaba clusters. Our study focuses on the characterization of microservice dependency as well as its runtime performance. We conduct an in-depth anatomy of microservice call graphs to quantify the difference between them and traditional DAGs of data-parallel jobs. In particular, we observe that microservice call graphs are heavy-tail distributed and their topology is similar to a tree and moreover, many microservices are hot-spots. We reveal three types of meaningful call dependency that can be utilized to optimize microservice designs. Our investigation on microservice runtime performance indicates most microservices are much more sensitive to CPU interference than memory interference. To synthesize more representative microservice traces, we build a mathematical model to simulate call graphs. Experimental results demonstrate our model can well preserve those graph properties observed from Alibaba traces.

Resource management is challenging in clouds due to the dynamics and sharing characteristics. The crucial problem is how to allocate resources accurately and satisfy demands of workloads timely. The traditional solution is to use historical data to predict future resource usage. Although these resource prediction methods can predict the periodicity, they can not accurately predict mutation points due to the high dynamics and uncertainty of resource usage. To tackle this issue, in this paper we propose a resource usage prediction method - RPTCN, which is based on a deep learning method - temporal convolutional networks (TCNs) in cloud systems. We add a fully connected layer and attention mechanism to TCNs to improve the prediction accuracy. In order to explore the relationship between the usage of different resources in the temporal dimension, we use correlation analysis to screen performance indicators as multidimensional feature input for prediction. Finally, we evaluate the performance of this method on Alibaba trace v2018. Evaluations show that RPTCN improves the overall MAE and MSE by 6.50%~89.03% and 0.41%~68.82% respectively compared to baselines in dynamic and long-term prediction of resource usage. Moreover, the convergence and generalization of RPTCN are also better than the baselines.

Face swapping has both positive applications such as entertainment, human-computer interaction, etc., and negative applications such as DeepFake threats to politics, economics, etc. Nevertheless, it is necessary to understand the scheme of advanced methods for high-quality face swapping and generate enough and representative face swapping images to train DeepFake detection algorithms. This paper proposes the first Megapixel level method for one shot Face Swapping (or MegaFS for short). Firstly, MegaFS organizes face representation hierarchically by the proposed Hierarchical Representation Face Encoder (HieRFE) in an extended latent space to maintain more facial details, rather than compressed representation in previous face swapping methods. Secondly, a carefully designed Face Transfer Module (FTM) is proposed to transfer the identity from a source image to the target by a non-linear trajectory without explicit feature disentanglement. Finally, the swapped faces can be synthesized by StyleGAN2 with the benefits of its training stability and powerful generative capability. Each part of MegaFS can be trained separately so the requirement of our model for GPU memory can be satisfied for megapixel face swapping. In summary, complete face representation, stable training, and limited memory usage are the three novel contributions to the success of our method. Extensive experiments demonstrate the superiority of MegaFS and the first megapixel level face swapping database is released for research on DeepFake detection and face image editing in the public domain.

While recent studies on semi-supervised learning have shown remarkable progress in leveraging both labeled and unlabeled data, most of them presume a basic setting of the model is randomly initialized. In this work, we consider semi-supervised learning and transfer learning jointly, leading to a more practical and competitive paradigm that can utilize both powerful pre-trained models from source domain as well as labeled/unlabeled data in the target domain. To better exploit the value of both pre-trained weights and unlabeled target examples, we introduce adaptive consistency regularization that consists of two complementary components: Adaptive Knowledge Consistency (AKC) on the examples between the source and target model, and Adaptive Representation Consistency (ARC) on the target model between labeled and unlabeled examples. Examples involved in the consistency regularization are adaptively selected according to their potential contributions to the target task. We conduct extensive experiments on popular benchmarks including CIFAR-10, CUB-200, and MURA, by fine-tuning the ImageNet pre-trained ResNet-50 model. Results show that our proposed adaptive consistency regularization outperforms state-of-the-art semi-supervised learning techniques such as Pseudo Label, Mean Teacher, and FixMatch. Moreover, our algorithm is orthogonal to existing methods and thus able to gain additional improvements on top of MixMatch and FixMatch. Our code is available at https://github.com/Walleclipse/Semi-Supervised-Transfer-Learning-Paddle.

Deep convolutional neural networks are susceptible to adversarial attacks. They can be easily deceived to give an incorrect output by adding a tiny perturbation to the input. This presents a great challenge in making CNNs robust against such attacks. An influx of new defense techniques have been proposed to this end. In this paper, we show that latent features in certain "robust" models are surprisingly susceptible to adversarial attacks. On top of this, we introduce a unified ℓ∞-norm white-box attack algorithm which harnesses latent features in its gradient descent steps, namely LAFEAT. We show that not only is it computationally much more efficient for successful attacks, but it is also a stronger adversary than the current state-of-the-art across a wide range of defense mechanisms. This suggests that model robustness could be contingent on the effective use of the defender’s hidden components, and it should no longer be viewed from a holistic perspective.

Cracks are irregular line structures that are of interest in many computer vision applications. Crack detection (e.g., from pavement images) is a challenging task due to intensity in-homogeneity, topology complexity, low contrast and noisy background. The overall crack detection accuracy can be significantly affected by the detection performance on fine-grained cracks. In this work, we propose a Crack Transformer network (CrackFormer) for fine-grained crack detection. The CrackFormer is composed of novel attention modules in a SegNet-like encoder-decoder architecture. Specifically, it consists of novel self-attention modules with 1x1 convolutional kernels for efficient contextual information extraction across feature-channels, and efficient positional embedding to capture large receptive field contextual information for long range interactions. It also introduces new scaling-attention modules to combine outputs from the corresponding encoder and decoder blocks to suppress non-semantic features and sharpen semantic ones. The CrackFormer is trained and evaluated on three classical crack datasets. The experimental results show that the CrackFormer achieves the Optimal Dataset Scale (ODS) values of 0.871, 0.877 and 0.881, respectively, on the three datasets and outperforms the state-of-the-art methods.

Federated learning is a novel machine learning framework that enables multiple devices to collaboratively train high-performance models while preserving data privacy. Federated learning is a kind of crowdsourcing computing, where a task publisher shares profit with workers to utilize their data and computing resources. Intuitively, devices have no interest to participate in training without rewards that match their expended resources. In addition, guarding against malicious workers is also essential because they may upload meaningless updates to get undeserving rewards or damage the global model. In order to effectively solve these problems, we propose FIFL, a fair incentive mechanism for federated learning. FIFL rewards workers fairly to attract reliable and efficient ones while punishing and eliminating the malicious ones based on a dynamic real-time worker assessment mechanism. We evaluate the effectiveness of FIFL through theoretical analysis and comprehensive experiments. The evaluation results show that FIFL fairly distributes rewards according to workers’ behaviour and quality. FIFL increases the system revenue by 0.2% to 3.4% in reliable federations compared with baselines. In the unreliable scenario containing attackers which destroy the model’s performance, the system revenue of FIFL outperforms the baselines by more than 46.7%.

Camera motion estimation, such as relative pose estimation and absolute pose estimation, are fundamental problems in computer vision and robotics. To obtain the motion parameters, classical methods rely on studying the properties of the geometric matrices, e.g., rotation matrix, essential matrix, homography matrix. The well known five-point algorithm was successfully derived using the singular constraint and trace constraints on the essential matrix. However, finding all the algebraic constraints is not always trivial for some recent problems. In this paper, we propose a simple and general technique to find complete algebraic constraints so that we can derive efficient algorithms. We show that using the quaternion to formulate the rotation matrix we can eliminate any unknowns from the original equations and obtain constraints on the rest of the unknowns based on Gröbner basis. We demonstrate that this approach can be applied to almost all the camera motion estimation and show its improvement compared to the existing methods. Further more, based on this elimination technique, we exploit new constraints for the relative pose estimation with gravity prior, and derive a new globally optimal algorithm to this problem. We compare our algorithm with the state-of-the-art methods on both synthetic and real-world data, and show the benefits including accuracy and efficiency.

A technological revolution is occurring in the field of robotics with the data-driven deep learning technology. However, building datasets for each local robot is laborious. Meanwhile, data islands between local robots make data unable to be utilized collaboratively. To address this issue, the work presents Peer-Assisted Robotic Learning (PARL) in robotics, which is inspired by the peer-assisted learning in cognitive psychology and pedagogy. PARL implements data collaboration with the framework of cloud robotic systems. Both data and models are shared by robots to the cloud after semantic computing and training locally. The cloud converges the data and performs augmentation, integration, and transferring. Finally, fine tune this larger shared dataset in the cloud to local robots. Furthermore, we propose the DAT Network (Data Augmentation and Transferring Network) to implement the data processing in PARL. DAT Network can realize the augmentation of data from multi-local robots. We conduct experiments on a simplified self-driving task for robots (cars). DAT Network has a significant improvement in the augmentation in self-driving scenarios. Along with this, the self-driving experimental results also demonstrate that PARL is capable of improving learning effects with data collaboration of local robots.

Federated Learning (FL) creates an ecosystem for multiple agents to collaborate on building models with data privacy consideration. The method for contribution measurement of each agent in the FL system is critical for fair credits allocation but few are proposed. In this paper, we develop a real-time contribution measurement method FedCM that is simple but powerful. The method defines the impact of each agent, comprehensively considers the current round and the previous round to obtain the contribution rate of each agent with attention aggregation. Moreover, FedCM updates contribution every round, which enable it to perform in real-time. Real-time is not considered by the existing approaches, but it is critical for FL systems to allocate computing power, communication resources, etc. Compared to the state-of-the-art method, the experimental results show that FedCM is more sensitive to data quantity and data quality under the premise of real-time. Furthermore, we developed federated learning open-source software based on FedCM. The software has been applied to identify COVID-19 based on medical images.

Coronavirus disease 2019 (COVID-19) has resulted in an ongoing pandemic. Since COVID-19 spreads mainly via close contact among people, social distancing has become an effective manner to slow down the spread. However, completely forbidding close contact can also lead to unacceptable damage to the society. Thus, a system that can effectively monitor people's social distance and generate corresponding alerts when a high infection probability is detected is in urgent need. In this paper, we propose SmartDistance, a smartphone based software framework that monitors people's interaction in an effective manner, and generates a reminder whenever the infection probability is high. Specifically, SmartDistance dynamically senses both the relative distance and orientation during social interaction with a well-designed relative positioning system. In addition, it recognizes different events (e.g., speaking, coughing) and determines the infection space through a droplet transmission model. With event recognition and relative positioning, SmartDistance effectively detects risky social interaction, generates an alert immediately, and records the relevant data for close contact reporting. We prototype SmartDistance on different Android smartphones, and the evaluation shows it reduces the false positive rate from 33% to 1% and the false negative rate from 5% to 3% in infection risk detection.

Data caching in the cloud is an efficient way to improve the QoS of diverse data applications. However, this benefit is not freely available, given monetary cost to manage the caches in the cloud. In this paper, we study the data caching problem in the cloud that is driven by the monetary cost reduction, instead of the hit rate under limited capacity as in traditional cases. In particular, given a stream of requestsRto a shared data item, we present a shortest-path based optimal algorithm that can minimize the total transfer and caching costs within O(mn) time for off-line case, here m represents the number of nodes in the network, while n is the length of the request stream. The cost model in this computation is semi-homo, which indicates that all pairs of nodes have the same transfer cost, but each cache server node has its own caching cost rate. Our off-line algorithm improves the previous results not only in reducing the time complexity from O(m2n) to O(mn), but also in relaxing the cost model to be semi-homogeneous, rendering the algorithm more practical in reality. Furthermore, we also study this problem in its online form, and by extending the anticipatory caching idea, we propose a 2-competitive online algorithm based on the same cost model and show its tightness by giving a lower bound of the competitive ratio as 2 - o(1) for any deterministic online algorithm. We provably achieve these results with our deep insights into the problem and careful analysis of the solution algorithms, together with a trace-based study to evaluate their performance in reality.

Hang Hua, Xingjian Li, Dejing Dou, Chengzhong Xu, Jiebo Luo. Proceedings of the 2021 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies. 2021.

Summary The amount of outsourced data grows rapidly. In recent years, cloud service providers integrate data deduplication systems with convergent encryption (CE) methods, in which a file encryption key is determined by its own content instead of the secret of a specific user, to save the storage cost and ensure the security of outsourced data. However, present secure deduplication systems failed to deal with data modifications efficiently. We observe that when a client makes small changes on an existing file, the current chunking algorithms cannot effectively detect the similarities and always create chunks with largely overlapped contents. It reduces data deduplication ratios and results in unnecessary overhead. In this paper, we propose Sed‐Dedup, an efficient secure delta encoding deduplication system to address this problem. In Sed‐Dedup, we introduce a novel delta encoding approach to store modified contents in delta files and leave the original files intact. Two schemes with different encoding policies are designed. Both of them can solve the issue and improve the secure deduplication performance. To evaluate the performance, we implement a prototype and conduct extensive experiments based on synthetic and real‐world datasets. Our experimental results show that Sed‐Dedup is superior to the state‐of‐the‐art secure deduplication systems.

Summary Traffic bottlenecks dynamically change with the variance of traffic demand. Identifying traffic bottlenecks plays an important role in traffic planning and provides decision making. However, traffic bottlenecks are difficult to identify because of the complexity of traffic road networks and many other factors. In this article, we propose an influence spreading based method to find the dynamic changed traffic bottlenecks, where the influence caused by bottlenecks is maximal. We first build a traffic congestion diffusion (TCD) model to capture traffic flow influence (TFI) spreading over traffic road networks. The bottlenecks identification problem based on TCD is modeled as an influence maximization problem, that is, selecting the most influential nodes such that the deterioration of traffic condition is maximal. With the proof of the submodularity of TFI spreading over traffic networks, a provably near‐optimal algorithm is used to solve the NP‐hard problem. With the exploration of unique properties of TFI spread, an approximate influence maximization method for TCD (TCD‐AIM) is proposed. To the best of our knowledge, this should be the first model for a metro‐city scale from the influence perspective. Experimental results show that TCD‐AIM finds bottlenecks with up to 130% congestion density increase in the future.

The operation and management of intelligent transportation systems (ITS), such as traffic monitoring, relies on real-time data aggregation of vehicular traffic information, including vehicular types (e.g., cars, trucks, and buses), in the critical roads and highways. While traditional approaches based on vehicular-embedded GPS sensors or camera networks would either invade drivers’ privacy or require high deployment cost, this article introduces a low-cost method, namely,SenseMag, to recognize the vehicular type using a pair of noninvasive magnetic sensors deployed on the straight road section.SenseMagfilters out noises and segments received magnetic signals by the exact time points that the vehicle arrives or departs from every sensor node. Furthermore,SenseMagadopts a hierarchical recognition model to first estimate the speed/velocity, then identify the length of the vehicle using the predicted speed, sampling cycles, and the distance between the sensor nodes. With the vehicle length identified and the temporal/spectral features extracted from the magnetic signals,SenseMagclassifies the types of vehicles accordingly. Some semiautomated learning techniques have been adopted for the design of filters, features, and the choice of hyperparameters. Extensive experiment based on real-word field deployment (on the highways in Shenzhen, China) shows thatSenseMagsignificantly outperforms the existing methods in both classification accuracy and the granularity of vehicle types (i.e., seven types bySenseMagversus four types by the existing work in comparisons). To be specific, our field experiment results validate thatSenseMagis with at least 90% vehicle type classification accuracy and less than 5% vehicle length classification error.

Two functions are essential and necessary for the wireless-powered communication network, which are energy beamforming and localization. On one hand, energy beamforming controls the wireless energy waves of the energy access point (E-AP) in order to activate the nodes for transmitting information. On the other hand, locating the nodes is important to network management and location-based services in the wireless power communication network (WPCN). For a large-scale network, cooperative localization that employs neighborhood nodes to participate in positioning unknown target nodes is highly accurate and efficient. However, how to use energy beamforming to achieve highly accurate localization is not fully investigated yet. In this article, we analyze the impacts of energy beamforming on the cooperative localization performance of WPCNs. We formulate the Fisher information matrix (FIM) and the corresponding Cramér-Rao lower bound (CRLB) for the full connected network and a single node, respectively. Then, we propose beamforming schemes to optimize the cooperative localization and the power consumption. For optimal localization problems, we derive the closed-form expression of the optimal energy beamforming. For the optimal energy efficiency problems, we propose semidefinite programming (SDP) solutions to achieve the minimum power consumption while using calibrations to approach the actual localization requirements. Further, we also analyze the impacts of channel uncertainty. Through extensive simulations, the results demonstrate the dominant factors of the localization performance, and the performance improvements of our proposed schemes, which outperform the existing power allocation schemes.

Ensemble learning is a widely used technique to train deep convolutional neural networks (CNNs) for improved robustness and accuracy. While existing algorithms usually first train multiple diversified networks and then assemble these networks as an aggregated classifier, we propose a novel learning paradigm, namely, “In-Network Ensemble” ( INE ) that incorporates the diversity of multiple models through training a SINGLE deep neural network. Specifically, INE segments the outputs of the CNN into multiple independent classifiers, where each classifier is further fine-tuned with better accuracy through a so-called diversified knowledge distillation process . We then aggregate the fine-tuned independent classifiers using an Averaging-and-Softmax operator to obtain the final ensemble classifier. Note that, in the supervised learning settings, INE starts the CNN training from random, while, under the transfer learning settings, it also could start with a pre-trained model to incorporate the knowledge learned from additional datasets. Extensive experiments have been done using eight large-scale real-world datasets, including CIFAR, ImageNet, and Stanford Cars, among others, as well as common deep network architectures such as VGG, ResNet, and Wide ResNet. We have evaluated the method under two tasks: supervised learning and transfer learning. The results show that INE outperforms the state-of-the-art algorithms for deep ensemble learning with improved accuracy.

Stochastic Gradient Langevin Dynamics (SGLD) have been widely used for Bayesian sampling from certain probability distributions, incorporating derivatives of the log-posterior. With the derivative evaluation of the log-posterior distribution, SGLD methods generate samples from the distribution through performing as a thermostats dynamics that traverses over gradient flows of the log-posterior with certainly controllable perturbation. Even when the density is not known, existing solutions still can first learn the kernel density models from the given datasets, then produce new samples using the SGLD over the kernel density derivatives. In this work, instead of exploring new samples from kernel spaces, a novel SGLD sampler, namely, Randomized Measurement Langevin Dynamics (RMLD) is proposed to sample the high-dimensional sparse representations from the spectral domain of a given dataset. Specifically, given a random measurement matrix for sparse coding, RMLD first derives a novel likelihood evaluator of the probability distribution from the loss function of LASSO, then samples from the high-dimensional distribution using stochastic Langevin dynamics with derivatives of the logarithm likelihood and Metropolis–Hastings sampling. In addition, new samples in low-dimensional measuring spaces can be regenerated using the sampled high-dimensional vectors and the measurement matrix. The algorithm analysis shows that RMLD indeed projects a given dataset into a high-dimensional Gaussian distribution with Laplacian prior, then draw new sparse representation from the dataset through performing SGLD over the distribution. Extensive experiments have been conducted to evaluate the proposed algorithm using real-world datasets. The performance comparisons on three real-world applications demonstrate the superior performance of RMLD beyond baseline methods.

In this paper, we extend the classic MaxOut strategy, originally designed for Multiple Layer Preceptors (MLPs), intoCOnvolutionalMaxOut (COMO) — a new strategy making deep convolutional neural networks wider with parameter efficiency. Compared to the existing solutions, such as ResNeXt for ResNet or Inception for VGG-alikes, COMO works well on both linear architectures and the ones with skipped connections and residual blocks. More specifically, COMO adopts a novelsplit-transform-mergeparadigm that extends the layers withspatial resolution reductioninto multiple parallel splits. For the layer with COMO, each split passes the input feature maps through a4D convolution operatorwith independentbatch normalization operatorsfor transformation, then merge into the aggregated output of the original sizes throughmax-pooling. Such a strategy is expected to tackle the potential classification accuracy degradation due to the spatial resolution reduction, by incorporating the multiple splits and max-pooling-based feature selection. Our experiment using a wide range of deep architectures shows that COMO can significantly improve the classification accuracy of ResNet/VGG-alike networks based on a large number of benchmark datasets. COMO further outperforms the existing solutions, e.g., Inceptions, ResNeXts, SE-ResNet, and Xception, that make networks wider, and it dominates in the comparison of accuracy versus parameter sizes.

Analyzing blockchain protocols is a notoriously difficult task due to the underlying large scale distributed networks. To address this problem, stochastic model-based approaches are often utilized. However, the abstract models in prior work turn out not to be adoptable to consortium blockchains as the consensus of such a blockchain often consists of multiple processes. To address the lack of efficient analysis tools, we propose a queueing network-based method for analyzing consistency properties of consortium blockchain protocols in this article. Our method provides a way to evaluate the performance of the main stages in blockchain consensus. We apply our framework to the Hyperledger Fabric system and recover key properties of the blockchain network. Using our method, we analyze the security properties of the ordering mechanism and the impact of delaying endorsement messages in consortium blockchain protocols. Then an upper bound is derived of the damage an attacker could cause who is capable of delaying the honest players' messages. Based on the proposed method, we employ analytical derivations to investigate both the security and performance features, and corroborate close agreement with measurements on a wide-area network testbed running the Hyperledger Fabric blockchain. With the proposed method, designers of future blockchains can provide a more rigorous analysis of their consortium blockchain schemes.

Scalability of distributed deep learning (DL) training with parameter server (PS) architecture is often communication constrained in large clusters. There are recent efforts that use a layer by layer strategy to overlap gradient communication with backward computation so as to reduce the impact of communication constraint on the scalability. However, the approaches could bring significant overhead in gradient communication. Meanwhile, they cannot be effectively applied to the overlap between parameter communication and forward computation. In this article, we propose and develop iPart, a novel approach that partitions communication and computation in various partition sizes to overlap gradient communication with backward computation and parameter communication with forward computation. iPart formulates the partitioning decision as an optimization problem and solves it based on a greedy algorithm to derive communication and computation partitions. We implement iPart in the open-source DL framework BigDL and perform evaluations with various DL workloads. Experimental results show that iPart improves the scalability of a cluster of 72 nodes by up to 94 percent over the default PS and 52 percent over the layer by layer strategy.

With the rise of network virtualization, the workloads deployed on data center are dramatically changed to support diverse service-oriented applications, which are in general characterized by the time-bounded service response that in turn puts great burden on the data-center networks. Although there have been numerous techniques proposed to optimize the virtual network allocation in data center, the research on coordinating them in a flexible and effective way to autonomically adapt to the workloads for service time reduction is few and far between. To address these issues, in this article we propose Sova, an autonomic framework that can combine the virtual dynamic SR-IOV (DSR-IOV) and the virtual machine live migration (VLM) for virtual network allocations in data centers. DSR-IOV is a SR-IOV-based virtual network allocation technology, but its operation scope is very limited to a single physical machine, which could lead to the local hotspot issue in the course of computation and communication, likely increasing the service response time. In contrast, VLM is an often-used virtualization technique to optimize global network traffic via VM migration. Sova exploits the software-defined approach to combine these two technologies with reducing the service response time as a goal. To realize the autonomic coordination, the architecture of Sova is designed based on the MAPE-K loop in autonomic computing. With this design, Sova can adaptively optimize the network allocation between different services by coordinating DSR-IOV and VLM in autonomic way, depending on the resource usages of physical servers and the network characteristics of VMs. To this end, Sova needs to monitor the network traffic as well as the workload characteristics in the cluster, whereby the network properties are derived on the fly to direct the coordination between these two technologies. Our experiments show that Sova can exploit the advantages of both techniques to match and even beat the better performance of each individual technology by adapting to the VM workload changes.

Cloud Computing paradigm has revolutionized IT industry and be able to offer computing as the fifth utility. With the pay-as-you-go model, cloud computing enables to offer the resources dynamically for customers anytime. Drawing the attention from both academia and industry, cloud computing is viewed as one of the backbones of the modern economy. However, the high energy consumption of cloud data centers contributes to high operational costs and carbon emission to the environment. Therefore, Green cloud computing is required to ensure energy efficiency and sustainability, which can be achieved via energy efficient techniques. One of the dominant approaches is to apply energy efficient algorithms to optimize resource usage and energy consumption. Currently, various virtual machine consolidation-based energy efficient algorithms have been proposed to reduce the energy of cloud computing environment. However, most of them are not compared comprehensively under the same scenario, and their performance is not evaluated with the same experimental settings. This makes users hard to select the appropriate algorithm for their objectives. To provide insights for existing energy efficient algorithms and help researchers to choose the most suitable algorithm, in this paper, we compare several state-of-the-art energy efficient algorithms in depth from multiple perspectives, including architecture, modelling and metrics. In addition, we also implement and evaluate these algorithms with the same experimental settings in CloudSim toolkit. The experimental results show the performance comparison of these algorithms with comprehensive results. Finally, detailed discussions of these algorithms are provided.

With ever increasing concerns on environmental issues caused by gasoline fuel based vehicles, electric vehicles (EVs) have attracted more and more attention from governments, industries, and customers [1] . The recent advancements in EVs have great potential to create a more environmentally friendly smart city. However, due to limited battery capacity, most current mainstream EVs still have quite limited driving range (e.g., 100 miles) [2] . How to ensure the continuous running of EVs on a large-scale road network (e.g., metropolitan city, interstate) becomes a major concern.

Fine-tuning the deep convolution neural network (CNN) using a pre-trained model helps transfer knowledge learned from larger datasets to the target task. While the accuracy could be largely improved even when the training dataset is small, the transfer learning outcome is similar with the pre-trained one with closed CNN weights[17], as the backpropagation here brings less updates to deeper CNN layers. In this work, we propose RIFLE - a simple yet effective strategy that deepens backpropagation in transfer learning settings, through periodically ReInitializing the Fully-connected LayEr with random scratch during the fine-tuning procedure. RIFLE brings significant perturbation to the backpropagation process and leads to deep CNN weights update, while the affects of perturbation can be easily converged throughout the overall learning procedure. The experiments show that the use of RIFLE significantly improves deep transfer learning accuracy on a wide range of datasets, outperforming known tricks for the similar purpose, such as dropout, dropconnect, stochastic depth, and cyclic learning rate, under the same settings with 0.5%-2% higher testing accuracy. Empirical cases and ablation studies further indicate RIFLE brings meaningful updates to deep CNN layers with accuracy improved.

The core network is the most basic facility in the entire telecommunications network, which is consists of large number of routers, switches and firewalls. Network management like re-planning routes or adjusting policies is very important to avoid failures. However, the timing of intervention is very challenging. Too early intervention will incur unnecessary overheads, and too late intervention will cause serious disaster. In this paper, we analyzed a large data set from a real-world core telecommunications network and proposed Linear Bayesian Neural Networks (LBNN)11Code available at https://github.com/YanyingLin/Lbnn to perceive the core network state changes and make decisions about network intervention. In particular, we considered three aspects of complexity, including the weight of the mutual effect between devices, the dependence on the time dimension of the network states, and the randomness of the network state changes. The entire model is extended to a probability model based on the Bayesian framework to better capture the randomness and variability of the data. Experimental results on real-world data set show that LBNN achieves very high detection accuracy, with an average of 92.1%.

With the improvement of data volume, computing power and algorithms, deep learning has achieved rapid development and showing excellent performance. Recently, many deep learning models are proposed to solve the problems in different areas. A recurrent neural network (RNN) is a class of artificial neural networks where connections between nodes form a directed graph along a temporal sequence. This allows it to exhibit temporal dynamic behavior, which makes it applicable to tasks such as handwriting recognition or speech recognition. However, the RNN relies heavily on the automatic learning ability to update parameters that concentrate on the data flow but seldom considers the feature extraction capability of the gate mechanism. In this paper, we propose a novel architecture to build the forget gate which is generated by multiple bases. Instead of using the traditional single-layer fully-connected network, we use a Multiple Attention (MA) based network to generate the forget gate which refines the optimization space of gate function and improve the granularity of the recurrent neural network to approximate the map in the ground truth. Due to the benefit of MA structure on the gate mechanism, the proposed MA-LSTM model achieves better feature extraction capability than other known models.

Stock price movement prediction is commonly accepted as a very challenging task due to the volatile nature of financial markets. Previous works typically predict the stock price mainly based on its own information, neglecting the cross effect among involved stocks. However, it is well known that an individual stock price is correlated with prices of other stocks in complex ways. To take the cross effect into consideration, we propose a deep learning framework, called Multi-GCGRU, which comprises graph convolutional network (GCN) and gated recurrent unit (GRU) to predict stock movement. Specifically, we first encode multiple relationships among stocks into graphs based on financial domain knowledge and utilize GCN to extract the cross effect based on these pre-defined graphs. To further get rid of prior knowledge, we explore an adaptive relationship learned by data automatically. The cross-correlation features produced by GCN are concatenated with historical records and then fed into GRU to model the temporal dependency of stock prices. Experiments on two stock indexes in China market show that our model outperforms other baselines. Note that our model is rather feasible to incorporate more effective stock relationships containing expert knowledge, as well as learn data-driven relationship.

Graph Convolutional Network (GCN) is adopted to tackle the problem of convolution operation in non-Euclidean space. Previous works on GCN have made some progress, however, one of their limitations is that the design of Adjacency Matrix (AM) as GCN input requires domain knowledge and such process is cumbersome, tedious and error-prone. In addition, entries of a fixed Adjacency Matrix are generally designed as binary values (i.e., ones and zeros) which can not reflect the real relationship between nodes. Meanwhile, many applications require a weighted and dynamic Adjacency Matrix instead of an unweighted and fixed AM, and there are few works focusing on designing a more flexible Adjacency Matrix. To that end, we propose an end-to-end algorithm to improve the GCN performance by focusing on the Adjacency Matrix. We first provide a calculation method callednodeinformationentropyto update the matrix. Then, we perform the search strategy in a continuous space and introduce the Deep Deterministic Policy Gradient (DDPG) method to overcome the drawback of the discrete space search. Finally, we integrate the GCN and reinforcement learning into an end-to-end framework. Our method can automatically define the Adjacency Matrix without prior knowledge. At the same time, the proposed approach can deal with any size of the matrix and provide a better AM for network. Four popular datasets are selected to evaluate the capability of our algorithm. The method in this paper achieves the state-of-the-art performance onCoraandPubmeddatasets, with the accuracy of 84.6% and 81.6% respectively.

Graph Neural Networks (GNNs) have attracted increasing attention due to the significant representation learning capacity for graphs. The traffic forecasting is a typical graph representation learning task, but it is challenging to model the complex spatial and temporal relationships in traffics. Traditional spectral approaches get filters based on the eigendecomposition, which depends on the Laplacian matrix of the graph. However, these approaches have expensive matrix operation on graph convolutions neural networks and are insufficient to tackle the spatial dependency. In this paper, we propose a novel graph neural network - Spatial-Temporal Multi-head Graph ATtention network (ST-MGAT), to deal with the traffic forecasting problem. We build convolutions on the graph directly. We consider the features of neighborhood nodes and the weights of the edges to generate new node representation. More specifically, there are two main modules: i) Temporal convolution blocks to capture the dynamic time correlations; ii) Graph attention networks to capture the dynamic spatial relations between nodes. Experimental results show that our model achieves up to 13% improvement over the state-of-the-art approaches in short-term, medium-term, and long-term highway traffic forecasting.

Subway passenger flow forecasting, an essential component of intelligent transportation system, is critical for traffic management, public safety, urban planning. However, it is very challenging due to the high nonlinearities and complex dynamic spatio-temporal dependencies of passenger flows. In this paper, we model the subway system as a directed weighted graph and propose a novel spatio-temporal deep learning framework, Multi-STGCnet, for forecasting short-term subway passenger flow at a station level. Specifically, Multi-STGCnet is mainly composed of two components, temporal component and spatial component. (1) The temporal component employs three long short-term memory network (LSTM)-based modules to capture three temporal properties of the target station, which are the interval closeness, daily periodicity, weekly trend. (2) The spatial component designs three spatial matrixes to extract spatial correlation of a target station with all other stations classified as near neighbors, middle neighbors and distant neighbors. Respectively, it adopts three graph convolution network (GCN) and LSTM combined modules to capture the spatio-temporal influences from different neighbors. Finally, the outputs of the two components are fused with different weights to generate prediction. We evaluate Multi-STGCnet on a real world dataset from the metro system in Shenzhen, China. Experiment results demonstrate that our model outperforms multiple baselines.

We propose an integrated approach to active exploration by exploiting the Cartographer method as the base SLAM module for submap creation and performing efficient frontier detection in the geometrically co-aligned submaps induced by graph optimization. We also carry out analysis on the reachability of frontiers and their clusters to ensure that the detected frontier can be reached by robot. Our method is tested on a mobile robot in real indoor scene to demonstrate the effectiveness and efficiency of our approach.

In this paper, a global descriptor for a LiDAR point cloud, called LiDAR Iris, is proposed for fast and accurate loop-closure detection. A binary signature image can be obtained for each point cloud after several LoG-Gabor filtering and thresholding operations on the LiDAR-Iris image representation. Given two point clouds, their similarities can be calculated as the Hamming distance of two corresponding binary signature images extracted from the two point clouds, respectively. Our LiDAR-Iris method can achieve a pose-invariant loop-closure detection at a descriptor level with the Fourier transform of the LiDAR-Iris representation if assuming a 3D (x,y,yaw) pose space, although our method can generally be applied to a 6D pose space by re-aligning point clouds with an additional IMU sensor. Experimental results on five road-scene sequences demonstrate its excellent performance in loop-closure detection.

Federated Learning is designed for multiple mobile devices to collaboratively train an artificial intelligence model while preserving data privacy. Instead of collecting the raw training data from mobile devices to the cloud, Federated Learning coordinates a group of devices to train a shared model in a distributed manner with the training data located on the devices. However, in order to effectively deploy Federated Learning on resource-constrained mobile devices, several critical issues including convergence rate, scalability and energy efficiency should be well addressed. In this paper, we propose MCFL, a multi-layer online coordination framework for high-performance energy efficient federated learning. MCFL consists of two layers: a macro-layer on the central server and a micro-layer on each participating device. In each training round, the macro coordinator performs two tasks, namely, selecting the right devices to participate, and estimating a time limit, such that the overall training time is significantly reduced while still guaranteeing the model accuracy. Unlike existing systems, MCFL removes the restriction that participating devices must be connected to power sources, thus allowing more timely and ubiquitous training. This clearly requires on-device training to be highly energy-efficient. To this end, the micro coordinator determines optimal schedules for hardware resources in order to meet the time limit set by the macro coordinator with the least amount of energy consumption. Tested on real devices as well as simulation testbed, MCFL has shown to be able to effectively balance the convergence rate, model accuracy and energy efficiency. Compared with existing systems, MCFL can achieve a speedup up to 8.66× and reduce energy consumption by up to 76.5% during the training process.

Previous passenger demand inference methods have insufficient accuracy because they fail to catch the influence of all random factors (e.g., weather, holiday). Also, existing taxicab dispatching methods are not directly applicable for electric taxicabs because they cannot optimize their charging. We present CD-Guide: an electric taxicab dispatching and charging approach based on customized training and Reinforcement Learning (RL). We studied a metropolitan-scale taxicab dataset, and found: histogram of passengers' origin buildings (i.e., where they come from) is useful for selecting suitable training data for inference model, passenger demand in different regions may be influenced by various unpredictable random factors, and taxicabs' charging time must be considered to avoid missing potential passengers. By saying suitable historical data, we mean the data that are under the influence of random factors similar as current time. Then, we develop a RL based method to guide a taxicab to maximize its probability of picking up a passenger, minimize the number of its missed passengers due to charging, and meanwhile avoid the taxicab from battery exhaustion. Our trace-driven experiments show that compared with previous methods, CD-Guide increases the total number of served passengers by 100%.

Previous Electric Vehicle (EV) charging scheduling methods and EV route planning methods require EVs to spend extra waiting time and driving burden for a recharge. With the advancement of dynamic wireless charging for EVs, Mobile Energy Disseminator (MED), which can charge an EV in motion, becomes available. However, existing wireless charging scheduling methods for wireless sensors, which are the most related works to the deployment of MEDs, are not directly applicable for the scheduling of MEDs on city-scale road networks. We present MobiCharger: a Mobile wireless Charger guidance system that determines the number of serving MEDs, and the optimal routes of the MEDs periodically (e.g., every 30 minutes). Through analyzing a metropolitan-scale vehicle mobility dataset, we found that most vehicles have routines, and the temporal change of the number of driving vehicles changes during different time slots, which means the number of MEDs should adaptively change as well. Then, we propose a Reinforcement Learning based method to determine the number and the driving route of serving MEDs. Our experiments driven by the dataset demonstrate that MobiCharger increases the medium state-of-charge and the number of charges of all EVs by 50% and 100%, respectively.

Mobile phones and tablets are becoming the primary platform of choice. However, these systems still suffer from limited battery and computation resources. A popular technique in mobile edge systems is computing outsourcing that augments the capabilities of mobile systems by migrating heavy workloads to resourceful clouds located at the edges of cellular networks. In the multi-site scenario, it is possible for mobile devices to save more time and energy by offloading to several cloud service providers. One of the most important challenges is how to choose servers to offload the jobs. In this paper, we consider a multi-site decision problem. We present a scheme to determine the proper assignment probabilities in a two-site mobile-edge computing system. We propose an open queueing network model for an offloading system with two servers and put forward performance metrics used for evaluating the system. Then in the specific scenario of a mobile chess game, where the data transmission is small but the computation jobs are relatively heavy, we conduct offloading experiments to obtain the model parameters. Given the parameters as arrival rates and service rates, we calculate the optimal probability to assign jobs to offload or locally execute and the optimal probabilities to choose different cloud servers. The analysis results confirm that our multi-site offloading scheme is beneficial in terms of response time and energy usage. In addition, sensitivity analysis has been conducted with respect to the system arrival rate to investigate wider implications of the change of parameter values.

Crowd sensing is an efficient way to collect heterogeneous information in the complicated infrastructures for fingerprinting based indoor localization. However, the information related to the dynamic trajectory are difficult to fuse due to the reliability issues from different devices and user moving habits. In this paper, we proposed a crowd sensing based indoor localization system with heterogeneous information fusion, which is called Wimage. Wimage can efficiently fuse multiple information sources related to location information, e.g., visual image, WiFi and geomagnetic data, even if the targets are moving with different and variable speeds. Then we design image-base subregion matching algorithm to locate the initial position and segmented weighted K-nearest neighbor algorithm to attain the matched trajectories in the database. A dynamic temporal warping algorithm is proposed for further calibrating the estimations. The experimental results indicate that with the helps from different kinds of information, the root mean square error is only below 0. 4m, which is highly accurate for locating a target in a large scale of indoor environment.

Magnetic MIMO is a wireless power transfer (WPT) system that employs multiple magnetic resonance coils to provide high efficient wireless power in the near field. Magnetic energy beamforming is a typical scheme to control the currents or voltages of the transmitter coils in order to achieve some objectives. Thus, the magnetic channel information is essential to magnetic beamforming (MagBF), and it needs complicated circuits and communication protocols to feedback such information. Such information may be not available due to the circuit limits or privacy concerns. In addition, the performance will be degraded with imperfect channel estimation in the noisy and mobile dynamic environment. In this case, only some limited feedback information is available, e.g., received power. In this article, we propose a random MagBF method to achieve maximum received power efficiency and simplify the system architecture. This scheme employs iterative Monte Carlo sampling and resampling to search an optimal beamforming solution based on the received power feedbacks. We design an online training protocol to implement the proposed scheme. It is computationally light and requires only limited feedback information, which avoids complex channel estimation or AC measurements. The simulation and real experimental results indicate that our algorithm can effectively increase the received power and approach the optimal performance with a fast convergent rate.

Computation offloading between the edge and the cloud is an effective way for deployed service to fully utilize the resources at both sides for its QoS improvement and overall cost reduction. Although the offloading problem has been intensively studied in the context of mobile computing, existing algorithms in most cases cannot be effectively migrated to the edge-cloud environment because their inter-partition communication costs are always deemed as symmetric, and their intra-partition communication costs are often ignored, which, though reasonable to the traditional case, are not valid to our settings anymore. In this article, we propose a new algorithmic approach to the offloading problem in the edge-cloud environment, where a heterogeneous model is advocated to incorporate the communication cost between co-resident tasks while considering the asymmetry of communication costs between non-coresident tasks. We prove the offloading problem with respect to this model is NP-hard, and thereby designing an efficient algorithm to obtain a sub-optimal solution. Additionally, we also show that in a homogeneous case when the intra-partition and inter-partition communication costs between any pair of interactive tasks are symmetric, an optimal offloading algorithm can be devised by transforming the problem into a classical min-cut problem. We implemented and evaluated the algorithms by offloading a PageRank-based application in a controlled edge-cloud setting. Our empirical results show that the proposed algorithm for the heterogeneous case is always efficient to find a better offloading scheme, compared with the selected existing algorithms, while for the homogeneous case, the proposed solution can efficiently achieve the optimal strategy.

Service migration between datacenters can reduce the network overhead within a cloud infrastructure; thereby, also improving the quality of service for the clients. Most of the algorithms in the literature assume that the client access pattern remains stable for a sufficiently long period so as to amortize such migrations. However, if such an assumption does not hold, these algorithms can take arbitrarily poor migration decisions that can substantially degrade system performance. In this paper, we approach the issue of performing service migrations for an unknown and dynamically changing client access pattern. We propose an online algorithm that minimizes the inter-datacenter network, taking into account the network load of migrating a service between two datacenters, as well as the fact that the client request pattern may change “quickly”, before such a migration is amortized. We provide a rigorous mathematical proof showing that the algorithm is 3.8-competitive for a cloud network structured as a tree of multiple datacenters. We briefly discuss how the algorithm can be modified to work on general graph networks with an O(log|V|) probabilistic approximation of the optimal algorithm. Finally, we present an experimental evaluation of the algorithm based on extensive simulations.

Service migration is an often-used approach in cloud computing to minimize the access cost by moving the service close to most users. Although it is effective in a certain sense, the service migration in existing research still suffers from some deficiencies in its evolutionary abilities in scalability , sensitivity , and adaptability to effectively react to the dynamically changing environments. This article proposes an evolutionary framework based on deep reinforcement learning for virtual service migration in large-scale mobile cloud centers. To enhance the spatio-temporal sensitivity of the algorithm, we design a scalable reward function for virtual service migration, redefine the input state, and add a Recurrent Neural Network ( RNN ) to the learning framework. Additionally, in order to enhance the adaptability of the algorithm, we also decompose the action space and exploit the network cost to adjust the number of virtual machine (VMs). The experimental results show that, compared with the existing results, the migration strategy generated by the algorithm can not only significantly reduce the total service cost and achieve the load balancing at the same time, but also address the burst situations with low cost in dynamic environments.

Stochastic Gradient Hamiltonian Monte Carlo (SGHMC) methods have been widely used to sample from certain probability distributions, incorporating (kernel) density derivatives and/or given datasets. Instead of exploring new samples from kernel spaces, this piece of work proposed a novel SGHMC sampler, namely Spectral Hamiltonian Monte Carlo (SpHMC), that produces the high dimensional sparse representations of given datasets through sparse sensing and SGHMC. Inspired by compressed sensing, we assume all given samples are low-dimensional measurements of certain high-dimensional sparse vectors, while a continuous probability distribution exists in such high-dimensional space. Specifically, given a dictionary for sparse coding, SpHMC first derives a novel likelihood evaluator of the probability distribution from the loss function of LASSO, then samples from the high-dimensional distribution using stochastic Langevin dynamics with derivatives of the logarithm likelihood and Metropolis–Hastings sampling. In addition, new samples in low-dimensional measuring spaces can be regenerated using the sampled high-dimensional vectors and the dictionary. Extensive experiments have been conducted to evaluate the proposed algorithm using real-world datasets. The performance comparisons on three real-world applications demonstrate the superior performance of SpHMC beyond baseline methods.

Traffic bottlenecks identification plays an important role in traffic planning and provides decision-making for prevention of traffic congestion. Although traffic bottlenecks widely exist, they are difficult to predict because of the changing traffic condition and traffic demand. In this paper, we introduce a traffic congestion diffusion (TCD) model with traffic flow influence (TFI) to capture the traffic dynamics and give a panoramic view for the city by cross domain data fusion. We proposed novel definition of bottleneck from the perspective of influence spread under TCD. The bottlenecks identification problem is modeled as an influence maximization problem, i.e., selecting the top K influential nodes in road networks under certain traffic conditions. We establish the submodularity of influence spread and solve the NP-hard optimal seed selection problem by using an efficient heuristic algorithm (TCD-IM) with provable near-optimal performance guarantees. To the best of our knowledge, this should be the first model for a metro-city scale from the influence perspective. The TCD-IM model is able to identify the dynamic traffic bottlenecks.

Computation offloading is an often-used optimization method that exploits servers with powerful and plentiful resources to maximize computation efficiency with minimum cost. In this method, a client application is usually modeled as a weighted directed acyclic graph (DAG), which is typically split into two distinct parts - one running on client device and the other on server machine. To simplify the model, the inter-part communication costs are always assumed to be symmetric and the intra-part communication costs are commonly ignored. Although these assumptions are reasonable to the offloading in traditional mobile computing, they are not valid anymore when considering the problem in the edgecloud environment, especially with the development of microservice, where a provisioned multi-machine cluster at each side is involved. To address this problem, we propose a new offloading model in this paper, where both the intra-part communication costs as well as the asymmetry of inter-part communication costs are incorporated to carry out the client application, which are not a part of previous approaches. Given this model, we first prove the offloading problem is NP-hard, then design an efficient greedy algorithm to obtain a sub-optimal solution. Our numerical results show that our algorithm for the new model is always efficient to find a better offloading scheme, compared with other existing algorithms that lack the notion of communication costs between tasks co-located at the same side and the asymmetry of communication costs crossing sides.

In this paper, we study the data caching problem in mobile cloud environment where multiple correlated data items could be packed and migrated to serve a predefined sequence of requests. By leveraging the spatial and temporal trajectory of requests, we propose a two-phase caching algorithm. We first investigate the correlation between data items to determine whether or not two data items could be packed to transfer, and then combine an existing dynamic programming (DP)-based algorithm and a greedy strategy to design a two-phase algorithm, named DP_Greedy, for effectively caching these shared data items to serve a predefined sequence of requests. Under homogeneous cost model, we prove the proposed algorithm is at most 2/α times worse than the optimal one in terms of the total service cost, where α is the defined discount factor, and also show that the algorithm can achieve this results within O(mn2) time and O(mn) space complexity for m caches to serve a n-length sequence. We evaluate our algorithm by effectively implementing it and comparing it with the non-packing case, the result show the proposed DP_Greedy algorithm not only presents excellent performances but is also more in line with the actual situation.

Modern in-memory data-intensive computing systems like Spark create long-lived containers to execute diverse types of applications. They rely on a cluster manager like YARN or Mesos to perform resource allocation to the containers. The cluster manager or scheduler requires users of the containers to reserve resources beforehand. It is a challenge to estimate just right amounts of memory to run the applications before execution, so as to avoid over-or under-provisioning of memory space. We discover a general property of memory reservation elasticity, which allows applications to run with a reservation limit smaller than they would ideally need while only paying a moderate performance penalty. Based on the property, we designed a system, namely MEER, which performs online estimation of minimum necessary amount of memory limit that achieves nearly optimal performance. We referred to it as optimal reservation, which divides memory over-provisioning from under-provisioning. It is non-trivial to efficiently estimate optimal reservations on line through one step without runtime history. MEER uses a two-step approach to dealing with the challenge: 1) Do robust profiling and probability density analysis of applications' memory footprints in two pilot runs. By using confidence levels for the prediction, we reduce the negative effects of container footprints' randomness and achieve a highly accurate online initial estimation (over 80% accuracy) of optimal reservation. 2) By exploiting a self-decay property of the analytical results, MEER adaptively performs iterative search based on a feed-back control mechanism over subsequent recurring executions. We implemented MEER atop of YARN and evaluated the prototype by running 15 benchmark workloads on a 16-node local cluster. Evaluation results show that it achieves an average accuracy of more than 95%. By deploying MEER on schedulers and allocating memory according to the optimal reservations, one could improve cluster memory utilization by about 40%. It reduces individual application execution time by 2 to 6 times on average compared to the state-of-the-art approaches. A 90 times peak speedup for PageRank in comparison with the default Spark/Yarn is observed.

Traditional alarm prediction in telecommunication networks mainly relies on expert knowledge. However, with the increasing complexity of telecommunication network, the traditional methods may not work well. It's necessary to study new automatic association rules extraction methods. In this paper, we proposed a method called DCSA (Density-based Clustering and Sequential Association Analysis) for alarm association rules mining. We use time density-based clustering and FP-Growth algorithm to mine the association rules in alarm data, which overcomes the drawbacks of sliding windows method. In addition, we design a sequential rules filtering module to eliminate the items that do not meet the sequential conditions in the original rules. Experiments on 7.5 million real alarm items from a telecommunication company of China show the sequential rules filtering module can greatly reduce the redundancy of association rules. We also demonstrate that the proposed DCSA method could effectively predict the occurrence of alarms.

Docker container is experiencing rapid development with the support from the industry like Google and Alibaba and is being widely used in large scale production cloud environment. For example, Alibaba has deployed millions of containers for its internal business, and most of the online services are already migrated to the containers. Those services are usually very complex, spanning multiple containers with complex interaction and dependency relationship. Detecting potential anomalies in such a large container-based cloud platform is very challenging. Traditional detection models usually use system resource metrics like CPU and memory usage, but rarely consider the relationship among components, causing high false positive rate. In this paper, we present a novel Anomaly Detection and root cause localization method based on Graph Similarity (ADGS) in the container-based cloud environment. We first monitor the response time and resource usage of each component in the application to determine whether the system status is normal or not. Then, we propose a new mechanism to locate the root cause of the anomalies based on graph similarity, investigating the anomaly propagation rules among cluster components. We implement and evaluate our method in a container-based environment. The results show that the proposed method can detect and determine the root cause of anomalies efficiently and accurately.

Virtual machine (VM) migration is a widely used technique in cloud computing systems to increase reliability. There are also many other reasons that a VM is migrated during its lifetime, such as reducing energy consumption, improving performance, maintenance, etc. During a live VM migration, the underlying VM continues being up until all or part of its data has been transmitted from source to destination. The remaining data are transmitted in an off-line manner by suspending the corresponding VM. The longer the off-line transmission time, the worse the performance of the respective VM. The above is because during the off-line data transmission, the VM service is down. Because a running VM's memory is subject to changes, already transmitted data pages may get dirtied and thus needing re-transmission. The decision of when suspending the VM is not a trivial task at all. The above is justified by the fact that when suspending the VM early we may result in transmitting off-line a significant amount of data degrading thus the VM's performance. On the other hand, a long waiting time to suspend the VM may result in re-transmitting a huge amount of dirty data, leading in that way to waste of resources. In this paper, we tackle the joint problem of minimizing both the total VM migration time (reflecting the resources spent during a migration) and the VM downtime (reflecting the performance degradation). The aforementioned objective functions are weighted according to the needs of the underlying cloud provider/user. To tackle the problem, we propose an online deterministic algorithm resulting in an strong competitive ratio, as well as a randomized online algorithm achieving significantly better results against the deterministic algorithm.

In this paper, we design an efficient deduplication algorithm based on the distributed storage architecture of Ceph. The algorithm uses on-line block-level data deduplication technology to complete data slicing, which neither affects the data storage process in Ceph nor alter other interfaces and functions in Ceph. Without relying on any central node, the algorithm maintains the characteristics of Ceph by designing a special hash object to store the data fingerprint, and uses the CRUSH algorithm to judge the data duplication based on calculation, instead of global search. The algorithm replaces the duplicate data with the deduplicated objects, which storage their fingerprints with less storage space. We compare the effects of different block sizes with respect to the performance and deduplication rates through experimental studies, and select the most appropriate block size in our prototype implementation. The experimental results show that the algorithm can not only effectively save the storage space but also improve the bandwidth utilization when reading and writing the duplicate data.

Deep convolutional neural networks (CNNs) are powerful tools for a wide range of vision tasks, but the enormous amount of memory and compute resources required by CNNs poses a challenge in deploying them on constrained devices. Existing compression techniques, while excelling at reducing model sizes, struggle to be computationally friendly. In this paper, we attend to the statistical properties of sparse CNNs and present focused quantization, a novel quantization strategy based on power-of-two values, which exploits the weight distributions after fine-grained pruning. The proposed method dynamically discovers the most effective numerical representation for weights in layers with varying sparsities, significantly reducing model sizes. Multiplications in quantized CNNs are replaced with much cheaper bit-shift operations for efficient inference. Coupled with lossless encoding, we build a compression pipeline that provides CNNs with high compression ratios (CR), low computation cost and minimal loss in accuracies. In ResNet-50, we achieved a 18.08x CR with only 0.24% loss in top-5 accuracy, outperforming existing compression methods. We fully compress a ResNet-18 and found that it is not only higher in CR and top-5 accuracy, but also more hardware efficient as it requires fewer logic gates to implement when compared to other state-of-the-art quantization methods assuming the same throughput.

Federated Learning is a technique for learning AI models through the collaboration of a large number of resourceconstrained mobile devices, while preserving data privacy. Instead of aggregating the training data from devices, Federated Learning uses multiple rounds of parameter aggregation to train a model, wherein the participating devices are coordinated to incrementally update a shared model with their own parameters locally learned. To efficiently deploy Federated Learning system over mobile devices, several critical issues including realtimeliness and energy efficiency should be well addressed. This paper proposes SmartPC, a hierarchical online pace control framework for Federated Learning that balances the training time and model accuracy in an energy-efficient manner. SmartPC consists of two layers of pace control: global and local. Prior to every training round, the global controller first oversees the status (e.g., connectivity, availability, and energy/resource remained) of every participating device, then selects qualified devices and assigns them a well-estimated virtual deadline for task completion. Within such virtual deadline, a statistically significant proportion (e.g., 60%) of the devices are expected to complete one round of their local training and model updates, while the overall progress of multi-round training procedure is kept up adaptively. On each device, a local pace controller then dynamically adjusts device settings such as CPU frequency so that the learning task is able to meet the deadline with the least amount of energy consumption. We performed extensive experiments to evaluate SmartPC on both Android smartphones and simulation platforms using well-known datasets. The experiment results show that SmartPC reduces up to 32:8% energy consumption on mobile devices and achieves a speedup of 2.27 in training time without model accuracy degradation.

Using IoT devices to locate targets is widely applied in many scenarios. However, replacing the batteries of these devices is time and labor consuming. In this article, we propose a wireless power-driven positioning system (WP2S) that employs MIMO-based wireless power transfer access points to supply energy to batteryless anchors. In this case, the IoT localization devices will have unlimited power. We formulate the equivalent Fisher information matrix (EFIM) as a fundamental tool to analyze the system performance. Then, we propose resource allocation schemes for optimal location estimation and energy efficiency problems by relaxing the objectives as semidefinite programming problems. In addition, we also analyze the impacts of channel uncertainty, anchor uncertainty, and NLOS for the performances of location estimation and energy consumption. The robust algorithms are developed according to uncertainty models. Both the analysis and simulations demonstrate that the estimation accuracy relies heavily on the transmitted power and the uncertainty models will consume more power to meet the location requirements.

In this paper we consider a new caching model that enables data sharing for network services in a cost-effective way. The proposed caching algorithms are characterized by using monetary cost and access information to control the cache replacements, instead of exploiting capacity-oriented strategies as in traditional approaches. In particular, given a stream of requests to a shared data item with respect to a homogeneous cost model, we first propose a fast off-line algorithm using dynamic programming techniques, which can generate an optimal schedule within$O(mn)$time-space complexity by using cache, migration as well as replication to serve a$n$-length request sequence in a$m$-node network, substantially improving the previous results. Furthermore, we also study the online form of this problem, and present an 3-competitive online algorithm by leveraging an idea of anticipatory caching. The algorithm can serve an online request in constant time and is space efficient in$O(m)$as well, rendering it more practical in reality. We evaluate our algorithms, together with some variants, by conducting extensive simulation studies. Our results show that the optimal cost of the off-line algorithm is changed in a parabolic form as the ratio of caching cost to transfer cost is increased, and the online algorithm is less than 2 times worse in most cases than its optimal off-line counterpart.

Reducing tail latency becomes increasingly important to improve user-perceived service experience. User-facing latency-sensitive cloud applications typically contain multiple interactive tiers running in different virtual machines (VMs) with complex interaction patterns. Consolidation of those applications is a challenge. In this paper we study the consolidation of multi-tier interactive workloads from a new perspective of user-perceived tail latency. We propose a novel profiling-based consolidation methodology. The objective is to satisfy tail latency while reducing the number of physical machines. We consider two key factors that affecting the tail latency of multi-tier workloads: interference with neighboring VMs and interaction between different tiers. We model the consolidation of multi-tier workloads as an optimization problem with different objectives and constraints. We implement and evaluate the proposed models, as well as comparing with other methods (i.e., without profiling or without considering interaction influence). Experimental results show that the proposed method is able to greatly reduce the tail latency compared with the traditional consolidation method.

Cloud computing technology is widely used in today's datacenters due to the benefits such as high scalability, on-demand services and low cost. An in-depth understanding of the characteristics of workloads running in production cloud environments is very important for improving the resource management efficiency. In this paper, we make a detailed analysis with visualization techniques and clustering methods on the trace dataset released by Alibaba which contains 11089 online services and 12951 batch jobs running on 1313 machines. Our methodology for clustering workloads contains: i) Select effective feature vectors as the dimensions of clustering; ii) Identify the cluster boundaries of each dimension using K-Means algorithm; iii) Classify jobs by combining the feature vectors which uses the results from previous step; iv) Analyze the characteristics of workload groups at runtime. Our analysis reveals several insights which previous work has not found on Alibaba cluster trace. For batch jobs: a) Average CPU cores of all batch jobs show bimodal-distribution obviously. b) At a random sampling time, more than 50 % machines only run one group of jobs with a short duration, medium CPU cores and small memory utilization, the remaining machines run mixed groups of jobs. For online instances: a) The resource usage (CPU, Memory, and Disk) of most online instances is low; b) There are up to six groups running on the same machine according to our clustering method at a random sampling time.

Minimizing service-level agreement (SLA) violations and energy consumption through server consolidation is of paramount importance for the sustainability of cloud environments. In this paper, we propose an online method to reduce cloud SLA violations by taking into account: (a) the energy consumption of migrating virtual machines and (b) server consolidation techniques to minimize the energy consumption within the system. Rigorous mathematical competitive analysis shows that the proposed method achieves a 2.4 competitive ratio against a cognitive adversary. Our solution is superior compared to the current state of the art algorithms, such as UP-VMC, KMI, MBFD, both theoretically (competitive ratios of other alternatives are unbounded) and empirically through simulations using CloudSim. More specifically, the competitive ratios of state of the art algorithms are unbounded and our proposed methodology reduces the VM migrations and energy consumption by 85% and 45%, respectively. The aforementioned improvement comes at an expense of a small increase in terms of SLA violations. The above results are achieved without the a priori knowledge of VM utilization patterns.

Docker container is experiencing a rapid development with the support from industry like Google and is being widely used in large scale production cloud environments. However the performance of applications running in Docker containers is still not clear due to the complex relationship between container resource allocation and application performance. In this paper, we first study the impact of key parameters in container resource allocation that affect the performance of containerized applications. Then, we present modeling techniques over CPU, memory and I/O resources to characterize the performance of applications running in containers. To address this multi-dimensional modeling problem, we propose three machine learning techniques, i.e. Linear Regression (LR), Support Vector Machine (SVM) and Artificial Neural Network (ANN). We implement and evaluate the modeling techniques for four complex benchmark workloads from Spark. Experimental results demonstrate the proposed models can achieve as low as 2.27% prediction error, with an average of 10.13% for most applications. Furthermore, the prediction accuracy of SVM and ANN models are substantially better than LR based approaches, with 48.13% and 29.30% improvement.

Parallel file systems (PFSs) are widely deployed to speed up the performance of high-performance computing (HPC) applications. In recent years, hybrid PFSs that consist of HDD-SSD servers, have attracted much attention in HPC community. However, existing data layout schemes do not well consider the characteristics of heterogeneous servers and heterogeneous access patterns, thus may experience considerable inefficiencies. In this study, we propose MHA, a migratory heterogeneity-aware data layout scheme to improve the data distribution of hybrid PFS. More specifically, to accommodate heterogeneous access patterns, MHA first migrates file data into several regions, each with similar access patterns. Then, by leveraging a data access cost model, MHA determines the appropriate stripe sizes on heterogeneous servers to get the best performance on each region. We have implemented MHA under MPI-IO library on top of OrangeFS file system. Experimental results show that MHA can significantly improve the hybrid PFS I/O system performance compared to existing data layout schemes.

The radio-frequency (RF) enabled wireless power transfer (WPT) system can be benefit from the MIMO technique. However, due to the limited resource, internet of things (IoT) devices can only feedback partial information which is received signal strength (RSS) value instead of channel state information (CSI). Thus, channel estimation based beamforming scheme from receiver side is not applicable for real applications. In this paper, we propose an adaptive random beamforming algorithm based on Monte-Carlo method to supply multiple batteryless IoT devices with high received power efficiency. Our algorithm does not require the complex channel estimation and adapts the beamforming scheme only according to the partial feedback information. We employ Gibbs sampling and re-sampling methods to generate several random beamforming weight vectors, and choose the optimal one. A simulated annealing algorithm is employed to control the convergence rate. We use the proposed algorithm to supply power in two cases: the maximum power transmission and robust power transmission. The simulation results indicate that this algorithm can fast converge to an optimal value and provide far-field power to multiple IoT devices.

Graphic processing units (GPUs) are widely used for general-purpose computing-so-called GPGPU computing. GPUs feature a large number of architecture parameters, resulting in a huge design space. To quickly explore this design space and identify the optimum architecture for a group of widely used computing kernels, it is critical to know how important each parameter is and how strongly these parameters interact with each other. This paper proposes an ensemble-learning-based approach, called quantifying the importance and interaction of Gpgpu architecture parameters (QIG), to quantify the importance of architecture parameters and their interactions with respect to performance. QIG employs a stochastic gradient boosted regression tree to construct performance models using performance data from a random set of GPU architectures. Leveraging these models, QIG observes the impact of each architecture parameter on performance, and calculates its importance and interaction intensity with other parameters. Using 25 widely used GPGPU kernels, we demonstrate that QIG accurately ranks the importance and interaction of GPU architecture parameters while the previously proposed Plackett-Burman design does not. Moreover, we show that QIG leads to a substantially more accurate performance model compared to prior work, including Starchart and approaches using artificial neural networks and supported vector machines: average error of 4.2% for QIG versus 23+% for prior work. Finally, QIG reveals a number of interesting insights for GPU architectures running GPGPU workloads.

With the popularity of cloud computing, mobile devices can store/retrieve personal data from anywhere at any time. Consequently, the data security problem in mobile cloud becomes more and more severe and prevents further development of mobile cloud. There are substantial studies that have been conducted to improve the cloud security. However, most of them are not applicable for mobile cloud since mobile devices only have limited computing resources and power. Solutions with low computational overhead are in great need for mobile cloud applications. In this paper, we propose a lightweight data sharing scheme (LDSS) for mobile cloud computing. It adopts CP-ABE, an access control technology used in normal cloud environment, but changes the structure of access control tree to make it suitable for mobile cloud environments. LDSS moves a large portion of the computational intensive access control tree transformation in CP-ABE from mobile devices to external proxy servers. Furthermore, to reduce the user revocation cost, it introduces attribute description fields to implement lazy-revocation, which is a thorny issue in program based CP-ABE systems. The experimental results show that LDSS can effectively reduce the overhead on the mobile device side when users are sharing data in mobile cloud environments.

Data analytics is at the foundation of both high-quality products and services in modern economies and societies. Big data workloads run on complex large-scale computing clusters, which implies significant challenges for deeply understanding and characterizing overall system performance. In general, performance is affected by many factors at multiple layers in the system stack, hence it is challenging to identify the key metrics when understanding big data workload performance. In this paper, we propose a novel workload characterization methodology using ensemble learning, called Metric Importance Analysis (MIA), to quantify the respective importance of workload metrics. By focusing on the most important metrics, MIA reduces the complexity of the analysis without losing information. Moreover, we develop the MIA-based Kiviat Plot (MKP) and Benchmark Similarity Matrix (BSM) which provide more insightful information than the traditional linkage clustering based dendrogram to visualize program behavior (dis)similarity. To demonstrate the applicability of MIA, we use it to characterize three big data benchmark suites: HiBench, CloudRank-D and SZTS. The results show that MIA is able to characterize complex big data workloads in a simple, intuitive manner, and reveal interesting insights. Moreover, through a case study, we demonstrate that tuning the configuration parameters related to the important metrics found by MIA results in higher performance improvements than through tuning the parameters related to the less important ones.

To improve resource efficiency and design intelligent scheduler for clouds, it is necessary to understand the workload characteristics and machine utilization in large-scale cloud data centers. In this paper, we perform a deep analysis on a newly released trace dataset by Alibaba in September 2017, consists of detail statistics of 11089 online service jobs and 12951 batch jobs co-locating on 1300 machines over 12 hours. To the best of our knowledge, this is one of the first work to analyze the Alibaba public trace. Our analysis reveals several important insights about different types of imbalance in the Alibaba cloud. Such imbalances exacerbate the complexity and challenge of cloud resource management, which might incur severe wastes of resources and low cluster utilization. 1) Spatial Imbalance: heterogeneous resource utilization across machines and workloads. 2) Temporal Imbalance: greatly time-varying resource usages per workload and machine. 3) Imbalanced proportion of multi-dimensional resources (CPU and memory) utilization per workload. 4) Imbalanced resource demands and runtime statistics (duration and task number) between online service and offline batch jobs. We argue accommodating such imbalances during resource allocation is critical to improve cluster efficiency, and will motivate the emergence of new resource managers and schedulers.

Traffic congestion is a spatio-temporal state of speeds beyond the capacity of road design and congestion may propagate through road networks. Characterizing the diffusion process is of great importance both in congestion relief and traffic condition prediction. Traffic congestion diffusion (TCD) in road networks can be observed, but literature lacks accurate models for characterizing the process. In this paper, we define a concept of Traffic Flow Influence (TFI) as a base for congestion diffusion. A TCD model is designed to characterize not only the traffic flow evolving process in time domain but also the propagation process of TFI through road networks in space domain. The model is for traffic networks in a city, which is divided into grids and each grid is modeled by traffic status of congested or smooth. Different from other diffusion models, the grid status depends on not only its current condition, but also the relative traffic flow from and to its neighbors. We use a gradient descent approach to quantify the traffic flow and TFI intensity of road networks. To the best of our knowledge, this should be the first model for a metro-city scale. The TCD model with TFI is able to predict grid status with an accuracy as high as 89%. Experimental results based on real-world taxi trajectory data in a metro-city show that the TCD approach performs best in comparison with its competitors.

Modern in-memory distributed computation frameworks like Spark adequately leverage memory resources to cache intermediate data across multi-stage tasks in pre-allocated worker processes, so as to speedup executions. They rely on a cluster resource manager like Yarn or Mesos to pre-reserve specific amount of CPU and memory for workers ahead of task scheduling. Since a worker is executed for an entire application and runs multiple batches of DAG tasks from multi-stages, its memory demands change over time [3].

In this paper we consider the data caching problem in next generation data services in the cloud, which is characterized by using monetary cost and access trajectory information to control cache replacements, instead of exploiting capacityoriented strategies as in traditional research. In particular, given a stream of requests to a shared data item with respect to a homogeneous cost model, we first propose a fast off-line algorithm using dynamic programming techniques. The proposed algorithm can generate optimal schedule within O(mn) timespace complexity to cache, migrate as well as replicate the shared data item to serve an n-length request sequence with minimum cost in a fully connected m-node network, substantially improving the previous results. Additionally, we also study this problem in its online form, and present a 3-competitive online algorithm by leveraging a speculative caching idea. The algorithm can serve an online request in constant time, and is space efficient in O(m) as well, rendering it to be more practical in reality. Our research complements the shortage of similar research in literature on this problem.

The future generation of transportation system will be featured by electrified public transportation. To fulfill metropolitan transit demands, electric vehicles (EVs) must be continuously operable without recharging downtime. Wireless Power Transfer (WPT) techniques for in-motion EV charging is a solution. It however brings up a challenge: how to deploy charging lanes in a metropolitan road network to minimize the deployment cost while enabling EVs' continuous operability. In this paper, we propose CatCharger, which is the first work that handles this challenge. From a metropolitan-scale dataset collected from multiple sources of vehicles, we observe the diversity of vehicle passing speed and daily visit frequency (called traffic attributes) at intersections (i.e., landmarks), which are important factors for charging lane deployment. To select landmarks for deployment, we first group landmarks with similar traffic attribute values using the entropy minimization clustering method, and choose better candidate landmarks from each group suitable for deployment. To determine the deployment locations from the candidate landmarks, we infer the expected vehicle residual energy at each landmark using a Kernel Density Estimator fed by the vehicles' mobility, and formulate and solve an optimization problem to minimize the total deployment cost while ensuring a certain level of expected residual energy of EVs at each landmark. Our trace-driven experiments demonstrate the superior performance of CatCharger over other methods.

Migrating service to certain vantage locations that are close to its clients can not only reduce the service access latency,but also minimize the network costs for its service provider. As such, this problem is particularly important for time-bounded services to achieve both enhanced QoS and cost effectiveness as well. However, the service migration is not free, coming at costs of bulk-data transfer and likely service disruption, as a result, increasing the overall service costs. To gain the benefits of service migration while minimizing service costs, in this paper, we leverage reinforecement learning (RL) methods to propose an efficient algorithm, called Mig- RL, for the service migration in a cloud environment. The Mig-RL utilizes an agent to learn the optimal policy that determines service migration status by using a typical RL algorithm, called Q-learning. Specifically, the agent learns from the historical access information to decide when and to where the service should be migrated, without requiring any prior information regarding the service accesses. Therefore, the agent can dynamically adapt to the environment and achieve online migration in real time. Experimental results on the real and synthesized access sequences from cloud networks show that Mig-RL can minimize the service costs, and in the meantime, improve the quality of service (QoS) by adapting to the changes of mobile access patterns.

With increasing demands on high-performance and large-capacity SSDs in the enterprise-scale storage, the concern about the inefficient use of the DRAM space in SSDs rises, especially for those using page-level FTL (Flash Translation Layer). In such an FTL, the address mapping scheme allows a logical page address (LPA) to be mapped to any physical page address (PPA) in the disk. Though it provides flexible address management and minimizes internal data movements, it requires a large address mapping table whose size is proportional to the capacity of the disk. With the increase of SSD's capacity, the table can be too large to be held entirely in the DRAM buffer of the SSD, causing constantly accessing to the flash for the address translation. This performance penalty due to the buffer misses is particularly high with workloads of weak access locality and large working sets. In this paper, we propose a space- efficient page- level FTL using hash functions in the address translation, named Hash-based Page- level FTL, or HP-FTL in short, to address the concern. HP-FTL trades mapping flexibility with limited performance impact for high space efficiency allowing the entire table to fit in the buffer and eliminating translation misses. The experiment results show that HP-FTL can provide up to 2.6X throughput compared to DFTL, a representative page-level FTL, using the same amount of DRAM for buffering the table. Meanwhile, HP-FTL reduces the mapping table size to about 25% of the table space required by page- level mapping schemes, including DFTL, without having any buffer misses.

To predicate the future with high accuracy is a holy grail in financial market. However, the volatility of chaotic financial market challenges new technologies from computer science to economic science all the time. Recently, Recurrent Neural Network (RNN) plays a new role in financial market prediction. However, results from RNN are restricted by sample size of training datasets, and show predication accuracy can hardly be guaranteed in a long term. On the other hand, Representative Pattern Discovery (RPD) is an effective way in long-term prediction while it is ineffective in short-term prediction. In this paper, we define a representative pattern for time series, and propose a fusion financial prediction strategy based on RNN and RPD. We take the advantages of both RNN and RPD, in the way that the proposed strategy is stateful to keep the short-term trend and it rectifies the predication by a time-dependent incremental factor in a long-term way. Compared with RNN and pattern discovery respectively, our experimental results demonstrate that our proposed strategy performs much better than that of others. It can increase the prediction accuracy by 6% on the basis of RNN at most, but at a cost of higher Mean Squared Error.

Using multi-antennas can improve the received energy efficiency of the radio-frequency (RF) enabled wireless power transfer (WPT) system. However, for the resource constrained internet of things (IoT) devices, only partial information which is received signal strength (RSS) value instead of channel state information (CSI) can be fed back. In this paper, we propose an adaptive random beamforming algorithm based on Monte-Carlo method to achieve the maximum received power efficiency. The proposed algorithm does not require any complicated channel estimation and it adapts the beamforming scheme only according to the RSS values. Gibbs sampling is used to generate the random beamforming weight vectors and re-sample them according to the feedback RSS values in an iterative manner. In addition, we employ a simulated annealing algorithm to control the convergence rate. The simulation results indicate that this algorithm can fast converge to an optimal value and achieve the maximum received power.

Many exist localization algorithms are unbiased estimators. However, the estimation performance presents biased feature in the real location systems. On the other hand, many biased location estimators show advantages that unbiased estimators can not achieve, e.g., robust to the noise, more accurate estimation and low complexity. In this paper, we propose a biased localization estimator and a hybrid Kalman filtering algorithm. The proposed algorithm is robust to the complicated environment with high accuracy. Both theoretical analysis and experimental evaluation indicate that the proposed algorithm outperform the unbiased optimal estimation methods.

Summary In this paper, we propose a simulation model to study real‐world replication workflows for cloud storage systems. With this model, we present three new methods to maximize the storage space usage during replica creation, and two novel QoS aware greedy algorithms for replica placement optimization. By using a simulation method, our algorithms are evaluated, through a comparison with the existing placement algorithms, to show that (i) a more evenly distributed replicas for a data set can be achieved by using round‐robin methods in replica creation phase and (ii) the two proposed greedy algorithms, namedGS_QoSandGS_QoS_C1, not only have more economical results than those from Chenet al., but also guarantee the QoS for clients. Copyright © 2016 John Wiley & Sons, Ltd.

We study the problem of dynamically migrating a service in the cloud to satisfy an online sequence of mobile batch-request demands in a cost-effective way. The service may have single or multiple replicas, each running on a virtual machine. As the origin of mobile accesses frequently changes over time, this problem is particularly important for time-bounded services to achieve enhanced Quality of Service and cost effectiveness. Moving the service closer to the client locations not only reduces the service access latency but also minimizes the network costs for service providers. However, these benefits are not free. The migration comes at a cost of bulk-data transfer and service disruption, and hence, increasing the overall service costs. To gain the benefits of service migration while minimizing the caused monetary costs, we propose an efficient search-based algorithm Dmig to migrate a single server, and then extend it as a scalable algorithm, called mDmig , to the multi-server situation, a more general case in the cloud. Both algorithms are fully distributed, symmetric, and characterized by the effective use of historical access information to conduct virtual migration so that the limitations of local search in the cost reduction can be overcome. To evaluate the algorithms, we compared them with some existing algorithms and an off-line algorithm. Our simulation results showed that the proposed algorithms exhibit better performance in service migration by adapting to the changes of mobile access patterns in a cost-effective way.

Collective I/O is a widely used middleware technique that exploits I/O access correlation among multiple processes to improve I/O system performance. However, most existing implementations of collective I/O strategies are designed and optimized for homogeneous I/O systems. In practice, the homogeneity assumptions do not hold in heterogeneous parallel I/O systems, which consist of multiple HDD and SSD-based servers and become increasingly promising. In this paper, we propose a heterogeneity-aware collective-I/O (HACIO) strategy to enhance the performance of conventional collective I/O operations. HACIO reorganizes the order of I/O requests for each aggregator with awareness of the storage performance of heterogeneous servers, so that the hardware of the systems can be better utilized. We have implemented HACIO in ROMIO, a widely used MPI-IO library. Experimental results show that HACIO can significantly increase the I/O throughputs of heterogeneous I/O systems.

Parallel file system (PFS) is commonly used in high-end computing systems. With the emergence of solid state drives (SSDs), hybrid PFS, which consists of both HDD and SSD servers, provides a practical I/O system solution for data-intensive applications. However, most existing data layout schemes are inefficient for hybrid PFS due to their unawareness of server heterogeneities and workload changes in different parts of a file. In this study, we propose a heterogeneity-aware region-level data layout scheme, HARL, to improve the data distribution of a hybrid PFS. HARL first divides a file into fine-grained, varying sized regions according to the workload features of an application, then determines appropriate file stripe sizes on servers for each region based on the performance of heterogeneous servers. Furthermore, to further improve the performance of a hybrid PFS, we propose a dynamic region-level layout scheme, HARL-D, which creates multiple replicas for each region and redirects file requests to the proper replicas with the lowest access costs at the runtime. Experimental results of representative benchmarks and a real application show that HARL can greatly improve I/O system performance, and demonstrate the advantages of HARL-D over HARL.

Bus service is the most important function of public transportation. Besides the major goal of carrying passengers around, providing a comfortable travel experience for passengers is also a key business consideration. To provide a comfortable travel experience, effective bus scheduling is essential. Traditional approaches are based on fixed timetables. The wide adoptions of smart card fare collection systems and GPS tracing systems in public transportation provide new opportunities for using the data-driven approaches to fit the demand of passengers. In this paper, we associate these two independent data sets to derive the passengers' origin and destination. As the data are real time, we build a system to forecast the passenger flow in real time. To the best of our knowledge, this is the first paper, which implements a system utilizing smart card data and GPS data to forecast the passenger flow in real time.

Metro systems have become one of the most important public transit services in cities. It is important to understand individual metro passengers' spatio-temporal travel patterns. More specifically, for a specific passenger: what are the temporal patterns? what are the spatial patterns? is there any relationship between the temporal and spatial patterns? are the passenger's travel patterns normal or special? Answering all these questions can help to improve metro services, such as evacuation policy making and marketing. Given a set of massive smart card data over a long period, how to effectively and systematically identify and understand the travel patterns of individual passengers in terms of space and time is a very challenging task. This paper proposes an effective data-mining procedure to better understand the travel patterns of individual metro passengers in Shenzhen, a modern and big city in China. First, we investigate the travel patterns in individual level and devise the method to retrieve them based on raw smart card transaction data, then use statistical-based and unsupervised clustering-based methods, to understand the hidden regularities and anomalies of the travel patterns. From a statistical-based point of view, we look into the passenger travel distribution patterns and find out the abnormal passengers based on the empirical knowledge. From unsupervised clustering point of view, we classify passengers in terms of the similarity of their travel patterns. To interpret the group behaviors, we also employ the bus transaction data. Moreover, the abnormal passengers are detected based on the clustering results. At last, we provide case studies and findings to demonstrate the effectiveness of the proposed scheme.

Metro systems play an important role in meeting the demand for urban transportation in large cities. The understanding of passenger route choice is critical for public transit management. The wide deployment of automated fare collection (AFC) systems opens up a new opportunity. However, only each trip's tap-in and tap-out time stamp and stations can be directly obtained from AFC system records; the train and route chosen by a passenger are unknown, information necessary to solve our problem. While existing methods work well in some specific situations, they hardly work for complicated situations. In this paper, we propose a solution that needs no additional equipment or human involvement than the AFC systems. We develop a probabilistic model that can estimate from empirical analysis how the passenger flows are dispatched to different routes and trains. We validate our approach using a large-scale data set collected from the Shenzhen Metro system. The measured results provide us with useful input when building the passenger path choice model.

Multi-tiered Parallel I/O systems that combine traditional HDDs with emerging SSDs mitigate the cost burden of SSDs while benefiting from their superior I/O performance. While a multi-tiered parallel I/O system is promising for data-intensive applications in high-performance (HPC) domains, placing data on each tier of the system to achieve high I/O performance remains a challenge. In this paper, we propose a cost-aware region-level (CARL) data placement scheme in multi-tiered parallel I/O systems. CARL divides a large file into several small regions, and then places regions on different types of servers based on region access costs. CARL includes a static policy S-CARL and a dynamic policy D-CARL. For applications whose I/O access patterns are completely known, S-CARL calculates the region costs within the entire workload duration, and uses a static data placement scheme to selectively place regions on the proper servers. To adapt to applications whose access patterns are unknown in advance, D-CARL uses a dynamic data placement scheme which migrates data among different servers within each time window. We have implemented CARL under MPI-IO library and OrangeFS parallel file system environment. Our evaluation with representative benchmarks and an application shows that CARL is both feasible and able to improve I/O performance significantly.

In this paper, we consider to improve scientific workflows in cloud environments where data transfers between tasks are performed via provisioned in-memory caching as a service, instead of relying entirely on slower disk-based file systems. However, this improvement is not free since services in the cloud are usually charged in a “pay-as-you-go” model. As a consequence, the workflow tenants have to estimate the amount of memory that they would like to pay. Given the intrinsic complexity of the workflows, it would be very hard to make an accurate prediction, which would lead to either oversubscription or undersubscription, resulting in unproductive spending or performance degradation. To address this problem, we propose a concept of minmax memory claim (MMC) to achieve cost-effective workflow computations in in-memory cloud computing environments. The minmax-memory claim is defined as the minimum amount of memory required to finish the workflow without compromising its maximum concurrency. With the concept of MMC, the workflow tenants can achieve the best performance via in-memory computing while minimizing the cost. In this paper, we present the procedure of how to find the MMCs for those workflows with arbitrary graphs in general and develop optimal efficient algorithms for some well-structured workflows in particular. To further show the values of this concept, we also implement these algorithms and apply them, through a simulation study, to improve deadlock resolutions in workflow-based workloads when memory resources are constrained.

We present a network I/O allocation framework, called Raccoon, for workload-aware VM scheduling algorithm to facilitate hybrid I/O workloads in virtual environments. Raccoon combines the strengths of paravirtual I/O and SR-IOV techniques to not only minimize the network latency, but also optimize the bandwidth utilization for workload-aware VM scheduling. In Raccoon, a limited number of VFs in SR-IOV are granted to I/O-intensive VMs while the paravirtual Network Interface Cards (vNICs) are allocated to other non-I/O-intensive VMs as the default resources. With this design, Raccoon provides latency reduction and bandwidth guarantee under the premise that I/O-intensive VMs will always be granted the VFs to facilitate their I/O operations. The types of workloads in each VM are identified at runtime by modified XenMon. By leveraging the ACPI Hotplug technique, Raccoon can adaptively plugin and plugout the SR-IOV VFs upon the changes of VM requirements so that an efficient I/O workload-aware VM scheduling algorithm can be implemented based on the bonding driver technique. The experimental results reveal that Raccoon can combine the benefits of para-virtual I/O and SR-IOV techniques to improve the overall performance of virtualized platforms with VMs that have diverse I/O workloads.

Recent advances in chip design and integration technologies have led to the development of Single-Chip Cloud computers which are a microcosm of cloud datacenters. Those computers are based on Network-on-Chip (NoC) architectures with deep memory hierarchies. Developing scheduling algorithms to reduce data access latency as well as energy consumption is a major challenge for such architectures. In this paper, we propose a set of algorithms to jointly address the problem of task scheduling and data allocation in a unified approach. Moreover, we present a feasible system model for NoC based multicores considering a three-level memory hierarchy that effectively captures the energy consumed by various elements of system including: processing cores, caches, and NoC subsystem. Simulation results show the superiority of proposed algorithms compared to two state-of-the-art algorithms found in the literature. The experimental results clearly indicate that algorithms performing data and task scheduling in a joint fashion are superior against techniques implementing task and data scheduling separately.

Several virtual machine (VM) placement algorithms have been proposed and studied in the literature with various scopes such as server consolidation or network cost minimization. In most cases, decisions on VM migrations are taken without factoring in directly the data access cost by VMs. In this paper, we investigate the use of data replication in conjunction with the VM assignment problem and target on developing algorithms that decide both on which data should be replicated where and which VM must be migrated so as to minimize the network overhead among traditional cloud and mobile cloud systems. We discuss both the un-capacitated case and the more realistic case whereby datacenters (for the traditional cloud case) and micro-datacenters (for the mobile cloud case) have limited storage and computing capacity. We propose an algorithm based on hyper-graph partitioning to solve the aforementioned problem in an optimal way regarding the unconstrained case and extend it to capture storage and computing capacity constraints. Experimental evaluation shows that the proposed algorithm yields up to 53 percent network overhead reduction when compared to state-of-the-art algorithms found in the literature.

Transaction logging and snapshotting are techniques used to deliver durability to the data in in-memory data stores. Absolute durability guarantees are delivered to a system by sequentially recording the transaction logs and snapshots to a non-volatile disk. Recent advancements in database restoration techniques have given rise to lock-free fuzzy snapshots. Still the transaction log that completes the fuzzy snapshots is not lock-free. In addition to locking, the major overhead behind the transaction logging technique is the bottleneck involved in storing the logs to a persistent but slower disk. This paper concentrates on implementing an in-memory transaction logging system with a lesser disk dependency. This logging system mainly targets the distributed in-memory data stores that are transaction replicated, eventually consistent and fault tolerant to crash failures. By making logging in-memory, the performance will be improved, but during the crash fails, the state may be lost. On recovery, we restore the current state partially from the locally available fuzzy snapshot and the remaining from the non-failed nodes in the distributed replica. ZooKeeper, a distributed data store that offers distributed coordination as its major service is used to implement and test our research. On average, a 30 times write performance improvement has been achieved with this approach guaranteeing sufficient durability in replicated mode.

We propose a new concept of minmax memory claim (MMC) to achieve cost-effective workflow computations in in-memory cloud computing environments. The minmax-memory claim is defined as the minimum amount of memory required to finish the workflow without compromising its maximum concurrency. With MMC, the workflow tenants can achieve the best performance via the maximum concurrency while minimizing the cost to use the memory resources. In this paper, we present the algorithms to find the MMC for workflow computation and evaluate its value by applying it to deadlock avoidance algorithms.

The future generation transportation system will be featured by electrified public transportation. To fulfill metropolitan transit demands, electric vehicles (EVs) must be continuously operable without recharging downtime. Wireless Power Transfer (WPT) techniques for in-motion EV charging is a solution [1], [2]. It however brings up a challenge: how to deploy charging lanes in a metropolitan road network to minimize the deployment cost while enabling EVs' continuous operability.

We study the problem of autonomous service migration in the cloud to satisfy an online sequence of mobile batch-request demands in a cost-effective way. As the origins of the mobile accesses frequently change over time, this problem is particularly important for time-bounded services to achieve enhanced QoS and cost effectiveness. Moving the service closer to its client locations not only reduces the service access latency but also minimizes the network costs for service providers. However, the migration comes at costs of bulk-data transfer and service disruption, as a result, increasing the overall service costs. To gain the benefits of service migration while minimizing the service costs, we propose an efficient search-based algorithm Dmig the service migration in an autonomous way. Compared with existing algorithms, the proposed algorithm is fully distributed, symmetric, and characterized by the effective use of historical access information to perform virtual migration that overcomes the limitation of traditional local search in cost reduction. To evaluate the algorithm, we compared it with some existing algorithms, and show that the proposed algorithm exhibits better performance by adapting to the changes of mobile access patterns in a cost effective way.

Traffic lights in urban area can significantly influence the efficiency and effectiveness of transportation. The real-time scheduling information of traffic lights is fundamentally important for many intelligent transportation applications, such as shortest-time navigation and green driving advisory. However, existing traffic light scheduling identification systems either entail dedicated infrastructures or depend on specialized traffic traces, which hinders the popularity and real world deployment. Differently, we propose to identify real-time traffic light scheduling by analyzing taxi traces that are widely accessible from taxi companies. The key idea is to exploit the periodicity in traffic patterns, which is directly affected by traffic lights. We also develop advanced algorithms to identify red/green lights duration and signal change time. We evaluate our solution using over one billion taxi records from Shenzhen, China. The evaluation results validate the effectiveness of our system.

Graphics processing units (GPUs) have recently evolved into popular accelerators for general-purpose parallel programs -- so-called GPGPU computing. Although programming models such as CUDA and OpenCL significantly improve GPGPU programmability, optimizing GPGPU programs is still far from trivial. Branch divergence is one of the root causes reducing GPGPU performance. Existing approaches are able to calculate the branch divergence rate but are unable to reveal how the branches diverge in a GPGPU program. In this paper, we propose the Thread Similarity Matrix (TSM) to visualize how branches diverge and in turn help find optimization opportunities. TSM contains an element for each pair of threads, representing the difference in code being executed by the pair of threads. The darker the element, the more similar the threads are, the lighter, the more dissimilar. TSM therefore allows GPGPU programmers to easily understand an application's branch divergence behavior and pinpoint performance anomalies. We present a case study to demonstrate how TSM can help optimize GPGPU programs: we improve the performance of a highly-optimized GPGPU kernel by 35% by reorganizing its thread organization to reduce its branch divergence rate.

Parallel GPGPU applications rely on barrier synchronization to align thread block activity. Few prior work has studied and characterized barrier synchronization within a thread block and its impact on performance. In this paper, we find that barriers cause substantial stall cycles in barrier-intensive GPGPU applications although GPGPUs employ lightweight hardware-support barriers. To help investigate the reasons, we define the execution between two adjacent barriers of a thread block as a warp-phase. We find that the execution progress within a warp-phase varies dramatically across warps, which we call warp-phase-divergence. While warp-phase-divergence may result from execution time disparity among warps due to differences in application code or input, and/or shared resource contention, we also pinpoint that warp-phase-divergence may result from warp scheduling.

In an IMANET, mobile users access both text and media web contents on the Internet through gateway nodes, with web access patterns, i.e., the Zipf-like distribution or the Stretched Exponential distribution. To reduce data access delay from the Internet, we consider the cache placement problem in cooperative caching, which is that each mobile node selects a subset of web contents to cache cooperatively in its limited cache so that total access cost is minimized. It has been proved NP-hard. We propose a solution named Adaptive Allocation Cooperative Caching (AACC), which adaptively divides the cache space of each node into two components: altruistic and selfish, according to detected data access patterns. AACC aims to find the optimal cache allocation solution to allocate appropriate cache spaces for two components in order to minimize total access cost. Given the Zipf-like access distribution, we find a near-optimal allocation solution to the cache placement problem. Simulation results show that AACC achieves much better performance than the existing best cooperative caching strategy in IMANETs in terms of average query delay, caching overheads, and query success ratio. In particular, AACC reduces caching overheads by 40% in average.

Smart cities have given a significant impetus to manage traffic and use transport networks in an intelligent way. For the above reason, intelligent transportation systems (ITSs) and location-based services (LBSs) have become an interesting research area over the last years. Due to the rapid increase of data volume within the transportation domain, cloud environment is of paramount importance for storing, accessing, handling, and processing such huge amounts of data. A large part of data within the transportation domain is produced in the form of Global Positioning System (GPS) data. Such a kind of data is usually infrequent and noisy and achieving the quality of real-time transport applications based on GPS is a difficult task. The map-matching process, which is responsible for the accurate alignment of observed GPS positions onto a road network, plays a pivotal role in many ITS applications. Regarding accuracy, the performance of a map-matching strategy is based on the shortest path between two consecutive observed GPS positions. On the other extreme, processing shortest path queries (SPQs) incurs high computational cost. Current map-matching techniques are approached with a fixed number of parameters, i.e., the number of candidate points ( N CP ) and error circle radius (ECR), which may lead to uncertainty when identifying road segments and either low-accurate results or a large number of SPQs. Moreover, due to the sampling error, GPS data with a high-sampling period (i.e., less than 10 s) typically contains extraneous datum, which also incurs an extra number of SPQs. Due to the high computation cost incurred by SPQs, current map-matching strategies are not suitable for real-time processing. In this paper, we propose real-time map-matching (called RT-MM), which is a fully adaptive map-matching strategy based on cloud to address the key challenge of SPQs in a map-matching process for real-time GPS trajectories. The evaluation of our approach against state-of-the-art approaches is performed through simulations based on both synthetic and real-world datasets.

Sampled microarchitectural simulation of single-threaded applications is mature technology for over a decade now. Sampling multithreaded applications, on the other hand, is much more complicated. Not until very recently have researchers proposed solutions for sampled simulation of multithreaded applications. Time-Based Sampling (TBS) samples multithreaded application execution based on time—not instructions as is typically done for single-threaded applications—yielding estimates for a multithreaded application’s execution time. In this article, we revisit and analyze previously proposed TBS approaches (periodic and cantor fractal based sampling), and we obtain a number of novel and surprising insights, such as (i) accurately estimating fast-forwarding IPC , that is, performance in-between sampling units, is more important than accurately estimating sample IPC , that is, performance within the sampling units; (ii) fast-forwarding IPC estimation accuracy is determined by both the sampling unit distribution and how to use the sampling units to predict fast-forwarding IPC; and (iii) cantor sampling is more accurate at small sampling unit sizes, whereas periodic is more accurate at large sampling unit sizes. These insights lead to the development of Two-level Hybrid Sampling (THS) , a novel sampling methodology for multithreaded applications that combines periodic sampling’s accuracy at large time scales (i.e., uniformly selecting coarse-grain sampling units across the entire program execution) with cantor sampling’s accuracy at small time scales (i.e., the ability to accurately predict fast-forwarding IPC in-between small sampling units). The clustered occurrence of small sampling units under cantor sampling also enables shortened warmup and thus enhanced simulation speed. Overall, THS achieves an average absolute execution time prediction error of 4% while yielding an average simulation speedup of 40 × compared to detailed simulation, which is both more accurate and faster than the current state-of-the-art. Case studies illustrate THS’ ability to accurately predict relative performance differences across the design space.

Electric vehicle (EV) taxis have been introduced into the public transportation systems to increase EV market penetration. Different from regular taxis that can refuel in minutes, EV taxis' recharging cycles can be as long as one hour. Due to the long cycle, the bad decision on the charging station, i.e., choosing one without empty charging piles, may lead to a long waiting time of more than an hour in the worst case. Therefore, choosing the right charging station is very important to reduce the overall waiting time. Considering that the waiting time can be a nonnegligible portion to the total work hours, the decision will naturally affect the revenue of individual EV taxis. The current practice of a taxi driver is to choose a station heuristically without a global knowledge. However, the heuristical choice can be a bad one that leads to more waiting time. Such cases can be easily observed in current collected taxi data in Shenzhen, China. Our analysis shows that there exists a large room for improvement in the extra waiting time as large as 30 min/driver. In this paper, we provide a real-time charging station recommendation system for EV taxis via large-scale GPS data mining. By combining each EV taxi's historical recharging events and real-time GPS trajectories, the current operational state of each taxi is predicted. Based on this information, for an EV taxi requesting a recommendation, we can recommend a charging station that leads to the minimal total time before its recharging starts. Extensive experiments verified that our predicted time is relatively accurate and can reduce the cost time of EV taxis by 50% in Shenzhen.

Hadoop is a widely-used implementation framework of the MapReduce programming model for large-scale data processing. Hadoop performance however is significantly affected by the settings of the Hadoop configuration parameters. Unfortunately, manually tuning these parameters is very time-consuming, if at all practical. This paper proposes an approach, called RFHOC, to automatically tune the Hadoop configuration parameters for optimized performance for a given application running on a given cluster. RFHOC constructs two ensembles of performance models using a random-forest approach for the map and reduce stage respectively. Leveraging these models, RFHOC employs a genetic algorithm to automatically search the Hadoop configuration space. The evaluation of RFHOC using five typical Hadoop programs, each with five different input data sets, shows that it achieves a performance speedup by a factor of 2.11$\times$on average and up to 7.4$\times$over the recently proposed cost-based optimization (CBO) approach. In addition, RFHOC's performance benefit increases with input data set size.

In this paper, we focus on strategy configurations of multiple users to make cloud service reservation. We consider the problem from a game theoretic perspective and formulate it into a non-cooperative game among the multiple cloud users, in which each user is informed with incomplete information of other users. For each user, we design a utility function which combines the net profit with time efficiency and try to maximize its value. We solve the problem by employing variational inequality (VI) theory and prove that there exists a Nash equilibrium solution set for the formulated game. Then, we propose an iterative proximal algorithm (IPA), which is designed to compute a Nash equilibrium solution. The convergence of the IPA algorithm is also analyzed and we find that it converges to a Nash equilibrium if several conditions are satisfied. Finally, we conduct some numerical calculations to verify our theoretical analysis. The experimental results show that our proposed IPA algorithm converges to a stable state very quickly and improves the utilities of all users to certain extent by configuring a proper request strategy.

Link instability and channel interference cause significant performance degradation to mobile ad hoc cognitive networks (MACNets). Existing work designs routing and assigns channels separately or does not consider mobility prediction and channel vacation to primary nodes. In this paper, we investigate how to jointly optimize route setup and channel assignment. Firstly, we propose an integrated data transmission cost (IDTC) to quantitatively measure the communication quality of links. This novel routing metric IDTC integratively considers 1) node mobility, 2) co-channel interference among primary and cognitive nodes, 3) relay workload on a specified channel, and 4) distance between the relay and the destination node. We, then, design channel assignment algorithms that completely avoid the interference with primary nodes and minimize the conflict to cognitive nodes. Finally, we propose a joint stable routing and channel assignment (J-SRCA) protocol based on mobility prediction for the network throughput maximization. In our J-SRCA, each link selected hop by hop is simultaneously assigned an interference-avoiding channel during a route setup. NS2-based simulation results demonstrate that our J-SRCA significantly improves various network performance, and the higher interference degree cognitive networks experience, the more improvement our J-SRCA will bring to the networks.

With an ever increasing amount of data generated by scientific experiments, social networks and mobile as well as wireless sensor networks, reducing resource consumption by big data applications becomes of paramount importance. Towards this end, filtering data close to the data sources is a common strategy in order to reduce network traffic. Assuming a network of nodes, each potentially generating data and a query in the form of a single operator to be applied in these data, the basic statement of the operator placement problem is: find the best node to place the operator so that the network traffic is minimized. In this paper we study the problem of placing a set of communicating operators exhibiting a tree structure over a tree network of nodes with capacity constraints. We take advantage of our previous work on unconstrained placement in order to develop a new approach enabling both single and group operator migrations using evictions of hosted operators if free space is required. To enhance their applicability, the algorithms work in a distributed asynchronous manner, requiring only minimal knowledge at each network node. Results from simulation experiments show that the proposed algorithms reduce considerably network overhead against their counterparts.

Agent-based middleware platforms for wireless sensor networks (WSNs) have received a lot of attention during the last years, due to their great flexibility in re-programming, monitoring, handling, and optimizing the application as well as the whole system. Even though many algorithms have been proposed for the dynamic placement of agents within the WSN, they do not take into account the coordination aspects of such migrations. This not only may result in slow convergence but also in perpetual oscillations of agent migrations, degrading application and system performance. In this paper, we propose full-coordination, semi-coordination, and non-coordination agent migration strategies, and evaluate their convergence and network overhead. We also provide proofs that convergence is guaranteed when dynamic agent placement algorithms adopt our proposed strategies. Our results show that the semi-coordination strategy is superior in terms of both network overhead and convergence rate.

Warm up is a crucial issue in sampled micro architectural simulation to avoid performance bias by constructing accurate states for micro-architectural structures before each sampling unit. Not until very recently have researchers proposed Time-Based Sampling (TBS) for the sampled simulation of multi-threaded applications. However, warm up in TBS is challenging and complicated, because (i) full functional warm up in TBS causes very high overhead, limiting overall simulation speed, (ii) traditional adaptive functional warm up for sampling single-threaded applications cannot be readily applied to TBS, and (iii) check pointing is inflexible (even invalid) due to the huge storage requirements and the variations across different runs for multi-threaded applications. In this work, we propose Shorter On-Line (SOL) warm up, which employs a two-stage strategy, using 'prime' warm up in the first stage, and an extended 'No-State-Loss (NSL)' method in the second stage. SOL is a single-pass, on-line warm up technique that addresses the warm up challenges posed in TBS in parallel simulators. SOL is highly accurate and efficient, providing a good trade-off between simulation accuracy and speed, and is easily deployed to different TBS techniques. For the PARSEC benchmarks on a simulated 8-core system, two state-of-the-art TBS techniques with SOL warm up provide a 7.2× and 37× simulation speedup over detailed simulation, respectively, compared to 3.1× and 4.5× under full warm up. SOL sacrifices only 0.3% in absolute execution time prediction accuracy on average.

Data analytics is at the core of the supply chain for both products and services in modern economies and societies. Big data workloads however, are placing unprecedented demands on computing technologies, calling for a deep understanding and characterization of these emerging workloads. In this paper, we propose Shen Zhen Transportation System (SZTS), a novel big data Hadoop benchmark suite comprised of real-life transportation analysis applications with real-life input data sets from Shenzhen in China. SZTS uniquely focuses on a specific and real-life application domain whereas other existing Hadoop benchmark suites, such as Hi Bench and Cloud Rank-D, consist of generic algorithms with synthetic inputs. We perform a cross-layer workload characterization at both the job and micro architecture level, revealing unique characteristics of SZTS compared to existing Hadoop benchmarks as well as general-purpose multi-core PARSEC benchmarks. We also study the sensitivity of workload behavior with respect to input data size, and propose a methodology for identifying representative input data sets.

Virtualizing Hadoop clusters provides many benefits, including rapid deployment, on-demand elasticity and secure multi-tenancy. However, a simple migration of Hadoop to a virtualized environment does not fully exploit these benefits. The dual role of a Hadoop worker, acting as both a compute node and a data node, makes it difficult to achieve efficient IO processing, maintain data locality, and exploit resource elasticity in the cloud. We find that decoupling per-node storage from its computation opens up opportunities for IO acceleration, locality improvement, and on-the-fly cluster resizing. To fully exploit these opportunities, we propose StoreApp, a shared storage appliance for virtual Hadoop worker nodes co-located on the same physical host. To completely separate storage from computation and prioritize IO processing, StoreApp pro-actively pushes intermediate data generated by map tasks to the storage node. StoreApp also implements late-binding task creation to take the advantage of prefetched data due to mis-aligned records. Experimental results show that StoreApp achieves up to 61% performance improvement compared to stock Hadoop and resizes the cluster to the (near) optimal degree of parallelism.

In this paper, we propose a self-boosted co-scheduling(SBCO) algorithm to reduce synchronization latency among consolidated virtual machines. Different from conventional co-scheduling which requires all runnable sibling vCPUs that are from the same VM to be scheduled at precisely the same time, SBCO reorders all these sibling vCPUs threads coarsely at the same level in their respective run queue, then schedules them at the same time window, and maintains global fairness between consolidated VMs. SBCO minimizes costly pCPU preemption and preserves the flexibility of the dynamic mapping between vCPUs and pCPUs. We have implemented SBCO in KVM and conducted comprehensive evaluations with various workloads. Results shows that SBCO is able to reduce the number of context switches significantly and achieve overall performance improve up to 10% compared with other competitors and improve up to 60% compared with the default scheduler.

As the fault frequency is increasing with the component count in modern and future computer systems, resilience becomes increasingly critical. Existing work on anomaly detection and fault prediction enables failure avoidance techniques to circumvent fault effects proactively. In addition, traditional fault tolerance techniques can be applied to handle faults reactively. Different types of faults may affect different components of a system and have various manifestations. They need to be treated differently. However, the existing fault handling techniques uniformly treat all faults without considering their types and distinct properties. In this paper, we present a differentiated fault remediation framework with action selection (DFRAS) which integrates both preventive and reactive remediation actions differentiated for different types of faults with their urgency requirements. We investigate four major types of faults and identify candidate remediation actions. We apply the urgency requirements as constraints for action selection. We propose formal performance models to quantify the wasted time of the candidate actions, and develop a decision making method to select the best actions that minimize the overall remediation cost. We have implemented a prototype of DFRAS and evaluated its performance by simulations and experiments. Simulation and experimental results show that the integrated fault remediation strategies can significantly reduce the remediation overhead. The developed DFRAS system is lightweight, making it feasible for online fault management in large-scale systems.

Graphics processing units (GPU), due to their massive computational power with up to thousands of concurrent threads and general-purpose GPU (GPGPU) programming models such as CUDA and OpenCL, have opened up new opportunities for speeding up general-purpose parallel applications. Unfortunately, pre-silicon architectural simulation of modern-day GPGPU architectures and workloads is extremely time-consuming. This paper addresses the GPGPU simulation challenge by proposing a framework, called GPGPU-MiniBench, for generating miniature, yet representative GPGPU workloads. GPGPU-MiniBench first summarizes the inherent execution behavior of existing GPGPU workloads in a profile. The central component in the profile is the Divergence Flow Statistics Graph (DFSG), which characterizes the dynamic control flow behavior including loops and branches of a GPGPU kernel. GPGPU-MiniBench generates a synthetic miniature GPGPU kernel that exhibits similar execution characteristics as the original workload, yet its execution time is much shorter thereby dramatically speeding up architectural simulation. Our experimental results show that GPGPU-MiniBench can speed up GPGPU architectural simulation by a factor of 49× on average and up to 589×, with an average IPC error of 4.7 percent across a broad set of GPGPU benchmarks from the CUDA SDK, Rodinia and Parboil benchmark suites. We also demonstrate the usefulness of GPGPU-MiniBench for driving GPU architecture exploration.

Operator placement plays a key role in reducing the aggregate network overhead within a wireless sensor network (WSN) to extend battery life and the longevity of the network. Consequently, optimal algorithms for the operator placement problem (OPP) are of paramount importance to WSN performance. Unfortunately, the OPP becomes NP-complete when capacity constraints on the WSN nodes are taken into account. There are many algorithms in the literature that tackle the OPP; however, most of them consider tree-structured query graphs without limitations regarding the operators hosted by the WSN nodes. Therefore, there is a need to propose sophisticated approaches such that the problem is solved in an effective fashion. In this paper, we propose a fully distributed approach that takes into account the WSN node capacity constraints. The proposed approach is thoroughly evaluated through simulations and the results reveal that the proposed approach is superior to several state-of-the-art algorithms, such as DRA, DBA, MCFA, dFNS, and GRAL* found in the literature.

In virtualization environments, static memory allocation for virtual machines (VMs) can lead to severe service level agreement (SLA) violations or inefficient use of memory. Dynamic memory allocation mechanisms such as ballooning and memory hotplug were proposed to handle the dynamics of memory demands. However, these mechanisms so far have not been quantitatively or comparatively studied. In this paper, we first develop a runtime system called U-tube, which provides a framework to adopt memory hotplug or ballooning for dynamic memory allocation. We then implement fine-grained memory hotplug in Xen. We demonstrate the effectiveness of U-tube for dynamic memory management through two case studies: dynamic memory balancing and memory overcommitment. With these two case studies, we make a quantitative comparison between memory hotplug and ballooning. The experiments show that there is no absolute winner for different scenarios. Our findings can be very useful for practitioners to choose the suitable dynamic memory management techniques in different scenarios.

After clients outsource their data to the cloud, they will lose physical control of their data. Many schemes are proposed for clients to verify the integrity of their data. This paper considers a complementary problem: When a client claims that the server has lost their data, how can we be sure that the client is correct and honest about the loss? It is possible that the client's meta data is corrupted or the client is lying in order to blackmail the server. In addition, most previous work relies on sequential indices. However, the indices bring significant overhead to bind an index to each block. We propose to replace sequential indices with much flexible non-sequential {\it coordinates}. The binding of coordinates to data blocks is performed through a Coordinate Merkle Hash Tree (CMHT). Based on CMHT, we can improve both the average and the worst-case update overhead by simplifying the updating algorithm.

Expanding our knowledge about human mobility is essential for building efficient wireless protocols and mobile applications. Previous human mobility studies have typically been built upon empirical single-source data (e.g., cellphone or transit data), which inevitably introduces a bias against residents not contributing this type of data, e.g., call detail records cannot be obtained from the residents without cellphone activities, and transit data cannot cover the residents who walk or ride private vehicles. To address this issue, we propose and implement a novel architecture mPat to explore human mobility using multi-source data. A reference implementation of mPat was developed at an unprecedented scale upon the urban infrastructures of Shenzhen, China. The novelty and uniqueness of mPat lie in its three layers: (i) a data feed layer consisting of real-time data feeds from 24 thousand vehicles, 16 million smart cards and 10 million cellphones; (ii) a mobility abstraction layer exploring the correlation and divergence among the multi-source data to analyze and infer human mobility; and (iii) an application layer to improve urban efficiency based on the human mobility findings of the study. The evaluation shows that mPat achieves a 75% inference accuracy, and that its real-world application reduces passenger travel time by 36%.

Recent embedded middleware platforms enable the structuring of an application as a set of collaborating agents deployed on various nodes of the underlying wireless sensor network (WSN). Of particular importance is the network cost incurred due to agent communication, which in turn depends on how the agents are placed within the WSN system. In this paper, we present two agent migration algorithms with the aim of minimizing the total network overhead. The first one takes independent single agent migration decisions, while the second one considers groups of agents for migration. Both algorithms work in a fully distributed fashion based on the knowledge available locally at each node, and can be used both for one-shot initial application deployment as well as for the continuous updating of agent placement. We also propose two methodologies to tackle the problem when WSN nodes have limited capacity. We show through theoretical analysis that one of our algorithms (called GRAL$\ast$) always results in an optimal placement, while for the rest of the algorithms, we derive approximation ratios pertaining to their performance. We evaluate the performance of our algorithms through a series of simulation experiments. Results show that group migration algorithms are superior compared to single agent migration algorithms with the performance difference reaching 34% for some settings.

Recently, big data has been evolved into a buzzword from academia to industry all over the world. Benchmarks are important tools for evaluating an IT system. However, benchmarking big data systems is much more challenging than ever before. First, big data systems are still in their infant stage and consequently they are not well understood. Second, big data systems are more complicated compared to previous systems such as a single node computing platform. While some researchers started to design benchmarks for big data systems, they do not consider the redundancy between their benchmarks. Moreover, they use artificial input data sets rather than real world data for their benchmarks. It is therefore unclear whether these benchmarks can be used to precisely evaluate the performance of big data systems. In this paper, we first analyze the redundancy among benchmarks from ICTBench, HiBench and typical workloads from real world applications: spatio-temporal data analysis for Shenzhen transportation system. Subsequently, we present an initial idea of a big data benchmark suite for spatio-temporal data. There are three findings in this work: (1) redundancy exists in these pioneering benchmark suites and some of them can be removed safely. (2) The workload behavior of trajectory data analysis applications is dramatically affected by their input data sets. (3) The benchmarks created for academic research cannot represent the cases of real world applications.

An increasing number of new multicore systems use the Non-Uniform Memory Access architecture due to its scalable memory performance. However, the complex interplay among data locality, contention on shared on-chip memory resources, and cross-node data sharing overhead, makes the delivery of an optimal and predictable program performance difficult. Virtualization further complicates the scheduling problem. Due to abstract and inaccurate mappings from virtual hardware to machine hardware, program and system-level optimizations are often not effective within virtual machines. We find that the penalty to access the “uncore” memory subsystem is an effective metric to predict program performance in NUMA multicore systems. Based on this metric, we add NUMA awareness to the virtual machine scheduling. We propose a Bias Random vCPU Migration (BRM) algorithm that dynamically migrates vCPUs to minimize the system-wide uncore penalty. We have implemented the scheme in the Xen virtual machine monitor. Experiment results on a two-way Intel NUMA multicore system with various workloads show that BRM is able to improve application performance by up to 31.7% compared with the default Xen credit scheduler. Moreover, BRM achieves predictable performance with, on average, no more than 2% runtime variations.

Recently, power consumption of radio access networks (RAN) has attracted a lot of attention since energy cost takes up a vast portion of operational expenditure (OPEX). Specifically, more than half of the energy is consumed by base stations (BS). Baseband Unit (BBU), which is responsible for baseband signal processing, is an important part of BS, and BBU consolidation in the form of a cluster is an promising solution to enhance energy efficiency. Although there are several existing work addressing similar ideas from theoretical perspective, there is still a gap between theoretical and practical solutions. Hence, by leveraging cloud computing and software-defined radio technologies, we develop a BBU cluster testbed, eBase, to consolidates multiple isolated BBUs into a virtualized BBU and then provide unified baseband resource pool for RAN. We present the architecture and implementation of eBase in detail. We also design a distributed resource management middleware to realize comprehensive resource management strategy for BBU cluster in order to improve energy efficiency. Through experiments, we find eBase can reduce total energy consumption by about 20% while satisfying about 95% QoS. These results not only demonstrate BBU cluster's significant capability to reduce energy cost and improve resource utilization, but also show eBase is a capable prototyping testbed for cloud radio access network research.

In this paper we tackle the problem of virtual machine (VM) placement onto physical servers to jointly optimize two objective functions. The first objective is to minimize the total energy spent within a cloud due to the servers that are commissioned to satisfy the computational demands of VMs. The second objective is to minimize the total network overhead incurred due to: (a) communicational dependencies between VMs, and (b) the VM migrations performed for the transition from an old assignment scheme to a new one. We study different methodologies for solving the aforementioned problem. The first approach is based on VM packing algorithms that optimize the above objective functions separately, reaching a single solution. The other approach is to tackle simultaneously the two optimization targets and define a set of non-dominating solutions. Performance evaluation using simulation experiments reveals interesting trade-offs between energy consumption and network load.

Dynamic resource configuration is crucial to the provisioning of service level agreements (SLAs) in cloud computing. Most of today's autonomic resource configuration approaches are designed to scale a single type of resource. A few works are able to partition multiple resources, but mainly to meet the requirement of throughput. Unlike throughput, however, response time behaves nonlinearly with respect to resources. Therefore, these approaches are hardly applicable to dynamic sharing of multi-resources for the provisioning of response time guarantee. Moreover, the optimization of resource efficiency and utilization has great significance to IaaS providers. We show theoretically and experimentally that resource optimization lies in balanced configuration of resources. In this paper, we propose a framework, BConf, for dynamic balanced configuration of multi-resources for the provisioning of response time guarantee in virtualized clusters. BConf employs an integrated MPC (model predictive control) and adaptive PI (proportional integral) control approach (IMAP). MPC is applied to actively balance multiple resources using a novel resource metric. For the performance prediction, a gray-box model is built on generic OS and hardware metrics in addition to resource actuators and performance. We find out that resource penalty is an effective metric to measure the imbalanced degree of a configuration. Using this metric and the model, BConf tunes resources in a balanced way by minimizing the resource penalty while satisfying the response time target. Adaptive PI is used to coordinate with MPC by narrowing the optimization space to a promising region. Within BConf framework, resources are coordinated during contention. Experimental results with mixed TPC-W and TPC-C benchmarks show that BConf reduces resource usages by about 50% and 30% for TPC-W and TPC-C respectively, improves stability by more than 35.6%, and has a much shorter settling time, in comparison with a representative partition approach. The advantages of BConf in resource coordination are also demonstrated.

Recently, graphics processing units (GPUs) have opened up new opportunities for speeding up general-purpose parallel applications due to their massive computational power and up to hundreds of thousands of threads enabled by programming models such as CUDA. However, due to the serial nature of existing micro-architecture simulators, these massively parallel architectures and workloads need to be simulated sequentially. As a result, simulating GPGPU architectures with typical benchmarks and input data sets is extremely time-consuming. This paper addresses the GPGPU architecture simulation challenge by generating miniature, yet representative GPGPU kernels. We first summarize the static characteristics of an existing GPGPU kernel in a profile, and analyze its dynamic behavior using the novel concept of the divergence flow statistics graph (DFSG). We subsequently use a GPGPU kernel synthesizing framework to generate a miniature proxy of the original kernel, which can reduce simulation time significantly. The key idea is to reduce the number of simulated instructions by decreasing per-thread iteration counts of loops. Our experimental results show that our approach can accelerate GPGPU architecture simulation by a factor of 88X on average and up to 589X with an average IPC relative error of 5.6%.

Recently, power consumption of radio access networks (RAN) has attracted a lot of attention since energy cost takes up a vast portion of operational expenditure (OPEX). Specifically, more than half of the energy is consumed by base stations (BS). Baseband Unit (BBU), which is responsible for baseband signal processing, is an important part of BS but still lacks effective power management solutions to improve energy efficiency. We present a power management middleware, GreenBase, to improve energy efficiency for BBUs. GreenBase consists of a software-based power metering tool, power and performance monitors as well as managers, an extensible power management policy library, and an energy efficiency analyzer. To evaluate this middleware, we construct a BBU testbed running video streaming applications based on software-defined radio platform GNU Radio/USRP. The experimental results validate GreenBase can achieve power and performance target with integrated power management policies. In particular, we find GreenBase can reduce total energy consumption by about 13% while satisfying target statistic QoS using a stochastic QoS controller.

SUMMARY Data centers are experiencing a remarkable growth in the number of interconnected servers. Being one of the foremost data center design concerns, network infrastructure plays a pivotal role in the initial capital investment and ascertaining the performance parameters for the data center. Legacy data center network (DCN) infrastructure lacks the inherent capability to meet the data centers growth trend and aggregate bandwidth demands. Deployment of even the highest‐end enterprise network equipment only delivers around 50% of the aggregate bandwidth at the edge of network. The vital challenges faced by the legacy DCN architecture trigger the need for new DCN architectures, to accommodate the growing demands of the ‘cloud computing’ paradigm. We have implemented and simulated the state of the art DCN models in this paper, namely: (a) legacy DCN architecture, (b) switch‐based, and (c) hybrid models, and compared their effectiveness by monitoring the network: (a) throughput and (b) average packet delay. The presented analysis may be perceived as a background benchmarking study for the further research on the simulation and implementation of the DCN‐customized topologies and customized addressing protocols in the large‐scale data centers. We have performed extensive simulations under various network traffic patterns to ascertain the strengths and inadequacies of the different DCN architectures. Moreover, we provide a firm foundation for further research and enhancement in DCN architectures. Copyright © 2012 John Wiley & Sons, Ltd.

Computer architects rely heavily on microarchitecture simulation to evaluate design alternatives. Unfortunately, cycle-accurate simulation is extremely slow, being at least 4 to 6 orders of magnitude slower than real hardware. This longstanding problem is further exacerbated in the multi-/many-core era, because single-threaded simulation performance has not improved much, while the design space has expanded substantially. Parallel simulation is a promising approach, yet does not completely solve the simulation challenge. Furthermore, existing sampling techniques, which are widely used for single-threaded applications, do not readily apply to multithreaded applications as thread interaction and synchronization must now be taken into account. This work presents PCantorSim , a novel Cantor set (a classic fractal)--based sampling scheme to accelerate parallel simulation of multithreaded applications. Through the use of the proposed methodology, only less than 5% of an application's execution time is simulated in detail. We have implemented our approach in Sniper (a parallel multicore simulator) and evaluated it by running the PARSEC benchmarks on a simulated 8-core system. The results show that PCantorSim increases simulation speed over detailed parallel simulation by a factor of 20×, on average, with an average absolute execution time prediction error of 5.3%.

Cloud elasticity allows dynamic resource provisioning in concert with actual application demands. Feedback control approaches have been applied with success to resource allocation in physical servers. However, cloud dynamics make the design of an accurate and stable resource controller challenging, especially when application-level performance is considered as the measured output. Application-level performance is highly dependent on the characteristics of workload and sensitive to cloud dynamics. To address these challenges, we extend a self-tuning fuzzy control (STFC) approach, originally developed for response time assurance in web servers to resource allocation in virtualized environments. We introduce mechanisms for adaptive output amplification and flexible rule selection in the STFC approach for better adaptability and stability. Based on the STFC, we further design a two-layer QoS provisioning framework, DynaQoS, that supports adaptive multi-objective resource allocation and service differentiation. We implement a prototype of DynaQoS on a Xen-based cloud testbed. Experimental results on representative server workloads show that STFC outperforms popular controllers such as Kalman filter, ARMA and, Adaptive PI in the control of CPU, memory, and disk bandwidth resources under both static and dynamic workloads. Further results with multiple control objectives and service classes demonstrate the effectiveness of DynaQoS in performance-power control and service differentiation.

Cloud computing has a key requirement for resource configuration in a real-time manner. In such virtualized environments, both virtual machines (VMs) and hosted applications need to be configured on-the-fly to adapt to system dynamics. The interplay between the layers of VMs and applications further complicates the problem of cloud configuration. Independent tuning of each aspect may not lead to optimal system wide performance. In this paper, we propose a framework, namely CoTuner, for coordinated configuration of VMs and resident applications. At the heart of the framework is a model-free hybrid reinforcement learning (RL) approach, which combines the advantages of Simplex method and RL method and is further enhanced by the use of system knowledge guided exploration policies. Experimental results on Xen-based virtualized environments with TPC-W and TPC-C benchmarks demonstrate that CoTuner is able to drive a virtual server cluster into an optimal or near-optimal configuration state on the fly, in response to the change of workload. It improves the systems throughput by more than 30 percent over independent tuning strategies. In comparison with the coordinated tuning strategies based on basic RL or Simplex algorithm, the hybrid RL algorithm gains 25 to 40 percent throughput improvement.

According to the pay-per-use model adopted in clouds, the more the resources consumed by an application running in a cloud computing environment, the greater the amount of money the owner of the corresponding application will be charged. Therefore, applying intelligent solutions to minimize the resource consumption is of great importance. Because centralized solutions are deemed unsuitable for large-distributed systems or large-scale applications, we propose a fully distributed algorithm (called DRA) to overcome the scalability issues. The aforementioned problem can be solved by identifying an assignment scheme between the interacting components of an application, such as processes and virtual machines, and the computing nodes of a cloud system, such that the total amount of resources consumed by the respective application is minimized. The decisions for the transition from one assignment scheme to another one are made in a dynamic way and based only on local information. It should be stressed that DRA achieves convergence and always results in the optimal solution. We also show, through an experimental evaluation, that DRA achieves up to 55% network cost reduction when compared to the most recent algorithm in the literature.

Cloud computing is becoming increasingly prevalent in recent years. It introduces an efficient way to achieve management flexibility and economic savings for distributed applications. To take advantage of computing and storage resources offered by cloud service providers, data owners must outsource their data onto public cloud servers which are not within their trusted domains. Therefore, the data security and privacy become a big concern. To prevent information disclosure, sensitive data has to be encrypted before uploading onto the cloud servers. This makes plain text keyword queries impossible. As the total amount of data stored in public clouds accumulates exponentially, it is very challenging to support efficient keyword based queries and rank the matching results on encrypted data. Most current works only consider single keyword queries without appropriate ranking schemes. The multi-keyword query problem was being considered only recently. MRSE [1] is one of the first research works to define and address the problem of effective yet secure ranked multi-keyword search over encrypted cloud data. However, the keyword dictionary used in MRSE is static and must be rebuilt when the number of keywords in the dictionary increases. It also has severe out-of-order problems in the matching results and does not take the keyword access frequencies into account, which greatly affects its usability. In this paper, we propose a novel approach, called MKQE, to address these issues. Only minor changes in the dictionary structure have to be done when extra keywords are introduced. We also introduce new trapdoor generation and scoring algorithms to make in-order query results. Furthermore, the keyword access frequency is considered so as to select an adequate matching file set. We conduct extensive simulations and the results prove that our approach performs much better than previous solutions.

With the development of mobile code frameworks for embedded systems, an application can be structured as a set of cooperating components (agents) that are placed on the nodes of the system in a flexible way. Reducing the network traffic caused by the application components is a crucial issue for the increase in the lifetime of wireless embedded systems, since it is widely known that the communication cost plays the most significant role in the energy consumption of embedded devices. To this end, most placement algorithms place or move an agent towards the center of gravity of the communication workload. However, if the target node does not have enough capacity, the attempt is usually aborted. In this paper, we introduce eviction-enabled algorithms that allow nodes to free capacity by forcing a locally hosted agent to move to another node, even at a loss, to accept a new and potentially more beneficial agent. To the best of our knowledge, this is the first time that agents are evicted by local hosts to enable beneficial agent migrations and eventually improve the total network cost. In this paper, we provide algorithms tackling the aforementioned problem in a fully distributed manner. We also present and discuss the results of extensive simulations, showing that eviction-enabled algorithms can outperform their counterparts by up to 300%.

NAND flash memory based Solid State Drive (SSD) is increasingly popular as one of the major non-volatile storage devices. Due to the superior performance and energy efficiency properties, it becomes an important complimentary device between the main memory and the traditional mechanical Hard Disk Drive (HDD). It is also anticipated to substitute HDD as the mainstream secondary storage. Today, flash memory is widely used in embedded systems, hand-held devices, personal computers and even enterprise computer systems. To access the data on the flash, a software component called Flash Translation Layer (FTL) has to be applied to convert the file system logical address into the corresponding physical address. FTL has great impacts on the system overall performance. Numerous FTL algorithms have been proposed in the past decade. DFTL is one of the most popular page-level address mapping FTL algorithms. It has been considered to have the best flexibility. However, it has extra mapping information I/O overhead and cannot always achieve the optimal performance. In this paper, we propose CAST, a novel and efficient pagelevel FTL algorithm to relieve this issue. CAST reserves a small portion of embedded SRAM to cache most recently accessed logical-physical address mapping information. Unlike DFTL, we use a compact packing methodology. Consecutive logical-physical page mapping information is represented with only a single entry. Thus, more address mapping information can be maintained in the caching table, and the cache hit rates can be increased. To improve the garbage collection efficiency, CAST maintains multiple current data blocks simultaneously. When a new data write request comes, the system can select an appropriate one to conduct the process based on the request issuer and/or logical address information. Our simulation results show that CAST outperforms DFTL under various workloads and it can reduce the number of erase operations and decrease the I/O response time significantly.

Resource discovery is critical to the usability and accessibility of grid computing systems. Distributed Hash Table (DHT) has been applied to grid systems as a distributed mechanism for providing scalable range-query and multi-attribute resource discovery. Multi-DHT-based approaches depend on multiple DHT networks with each network responsible for a single attribute. Single-DHT-based approaches keep the resource information of all attributes in a single node. Both classes of approaches lead to high overhead. In this paper, we propose a Low-Overhead Range-query Multi-attribute (LORM) DHT-based resource discovery approach. Unlike other DHT-based approaches, LORM relies on a single compound graph-based DHT network and distributes resource information among nodes in balance by taking advantage of the compound graph structure. Moreover, it has high capability to handle the large-scale and dynamic characteristics of resources in grids. Experimental results demonstrate the efficiency of LORM in comparison with other resource discovery approaches. LORM dramatically reduces maintenance and resource discovery overhead. In addition, it yields significant improvements in resource location efficiency. We also analyze the performance of the LORM approach rigorously by comparing it with other multi-DHT-based and single-DHT-based approaches with respect to their overhead and efficiency. The analytical results are consistent with experimental results, and prove the superiority of the LORM approach in theory.

In-network data aggregation is a useful technique to reduce redundant data and to improve communication efficiency. Traditional data aggregation schemes for wireless sensor networks usually rely on a fixed routing structure to ensure data can be aggregated at certain sensor nodes. However, they cannot be applied in highly mobile vehicular environments. In this paper, we propose an adaptive forwarding delay control scheme, namely Catch-Up, which dynamically changes the forwarding speed of nearby reports so that they have a better chance to meet each other and be aggregated together. The Catch-Up scheme is designed based on a distributed learning algorithm. Each vehicle learns from local observations and chooses a delay based on learning results. The simulation results demonstrate that our scheme can efficiently reduce the number of redundant reports and achieve a good trade-off between delay and communication overhead.

Currently, virtualization technology has been widely adopted by cloud service providers to provide flexible and cost-effective resource sharing among users. On cloud platforms, computing resources are allocated on-demand dynamically and the application hosted on a virtual machine(VM) usually has the illusion of complete control of resources. Thus, a selfish VM may strategically compete for resource with other VMs to maximize its own benefit while at the cost of overall system performance. This problem poses new challenges to cloud providers, who must thwart non-cooperative behavior as well as allocating resource among selfish VMs efficiently. In this paper, we propose to utilize mechanism design to allocate resource among selfish VMs in a non-cooperative cloud environment. Because the accurate relationship between VM's valuation function and allocated resource may not be available in practice and the valuation function parameters may not noise free, we also propose to apply stochastic approximation methods to get stochastic solution for allocation and payment outcomes. We show through theoretical analysis and simulations that the proposed stochastic mechanism is efficient and incentive compatible. That is, the incorporation of mechanism design for virtualized resource allocation is able to enforce cooperation and achieve efficient resource utilization among selfish VMs in non-cooperative cloud systems.

Live migration of virtual machine (VM) provides a significant benefit for virtual server mobility without disrupting service. It is widely used for system management in virtualized data centers. However, migration costs may vary significantly for different workloads due to the variety of VM configurations and workload characteristics. To take into account the migration overhead in migration decision-making, we investigate design methodologies to quantitatively predict the migration performance and energy cost. We thoroughly analyze the key parameters that affect the migration cost from theory to practice. We construct two application-oblivious models for the cost prediction by using learned knowledge about the workloads at the hypervisor (also called VMM) level. This should be the first kind of work to estimate VM live migration cost in terms of both performance and energy in a quantitative approach. We evaluate the models using five representative workloads on a Xen virtualized environment. Experimental results show that the refined model yields higher than 90% prediction accuracy in comparison with measured cost. Model-guided decisions can significantly reduce the migration cost by more than 72.9% at an energy saving of 73.6%.

Cloud elasticity allows dynamic resource provisioning in concert with actual application demands. Feedback control approaches have been applied with success to resource allocation in physical servers. However, cloud dynamics make the design of an accurate and stable resource controller more challenging, especially when response time is considered as the measured output. Response time is highly dependent on the characteristics of workload and sensitive to cloud dynamics. To address the challenges, we extend a self-tuning fuzzy control (STFC) approach, originally developed for response time assurance in web servers to resource allocation in virtualized environments. We introduce mechanisms for adaptive output amplification and flexible rule selection in the STFC approach for better adaptability and stability. Based on the STFC, we further design a two-layer QoS provisioning framework, DynaQoS, that supports adaptive multi-objective resource allocation and service differentiation. We implement a prototype of DynaQoS on a Xen-based cloud testbed. Experimental results on an E-Commerce benchmark show that STFC outperforms popular controllers such as Kalman filter, ARMA and adaptive PI by at least 16% and 37% under both static and dynamic workloads, respectively. Further results with multiple control objectives and service classes demonstrate the effectiveness of DynaQoS in performance-power control and service differentiation.

Cloud computing has a key requirement for resource configuration in a real-time manner. In such virtualized environments, both virtual machines (VMs) and hosted applications need to be configured on-the-fly to adapt to system dynamics. The interplay between the layers of VMs and applications further complicates the problem of cloud configuration. Independent tuning of each aspect may not lead to optimal system wide performance. In this paper, we propose a framework, namely CoTuner, for coordinated configuration of VMs and resident applications. At the heart of the framework is a model-free hybrid reinforcement learning (RL) approach, which combines the advantages of Simplex and RL methods and is further enhanced by the use of system knowledge guided exploration policies. Experimental results on Xen-based virtualized environments with TPC-W and TPC-C benchmarks demonstrate that CoTuner is able to drive a virtual server system into an optimal or near optimal configuration state dynamically, in response to the change of workload. It improves the systems throughput by more than 30% over independent tuning strategies. In comparison with the coordinated tuning strategies based solely on Simplex or basic RL algorithm, the hybrid RL algorithm gains 30% to 40% throughput improvement. Moreover, the algorithm is able to reduce SLA violation of the applications by more than 80%.

Although cloud computing has gained sufficient popularity recently, there are still some key impediments to enterprise adoption. Cloud management is one of the top challenges. The ability of on-the-fly partitioning hardware resources into virtual machine(VM) instances facilitates elastic computing environment to users. But the extra layer of resource virtualization poses challenges on effective cloud management. The factors of time-varying user demand, complicated interplay between co-hosted VMs and the arbitrary deployment of multitier applications make it difficult for administrators to plan good VM configurations. In this paper, we propose a distributed learning mechanism that facilitates self-adaptive virtual machines resource provisioning. We treat cloud resource allocation as a distributed learning task, in which each VM being a highly autonomous agent submits resource requests according to its own benefit. The mechanism evaluates the requests and replies with feedback. We develop a reinforcement learning algorithm with a highly efficient representation of experiences as the heart of the VM side learning engine. We prototype the mechanism and the distributed learning algorithm in an iBalloon system. Experiment results on an Xen-based cloud test bed demonstrate the effectiveness of iBalloon. The distributed VM agents are able to reach near-optimal configuration decisions in 7 iteration step sat no more than 5% performance cost. Most importantly, iBalloon shows good scalability on resource allocation by scaling to 128 correlated VMs.

In this paper, we propose a distributed learning mechanism that facilitates self-adaptive virtual machines resource provisioning. We treat cloud resource allocation as a distributed learning task, in which each VM being a highly autonomous agent submits resource requests according to its own benefit. The mechanism evaluates the requests and replies with feedback. We develop a reinforcement learning algorithm with a highly efficient representation of experiences as the heart of the VM side learning engine. We prototype the mechanism and the distributed learning algorithm in an iBalloon system. Experiment results on a Xen-based cloud testbed demonstrate the effectiveness of iBalloon.

Server virtualization technology facilitates the creation of an elastic computing infrastructure on demand. There are cloud applications like server-based computing and virtual desktop that concern startup latency and require impromptu requests for VM creation in a real-time manner. Conventional template-based VM creation is a time consuming process and lacks flexibility for the deployment of stateful VMs. In this paper, we present an abstraction of VM substrate to represent generic VM instances in miniature. Unlike templates that are stored as an image file in disk, VM substrates are docked in memory in a designated VM pool. They can be activated into stateful VMs without machine booting and application initialization. The abstraction leverages an arrange of techniques, including VM miniaturization, generalization, clone and migration, storage copy-on-write, and on-the-fly resource configuration, for rapid deployment of VMs and VM clusters on demand. We implement a prototype on a Xen platform and show that a server with typical configuration of TB disk and GB memory can accommodate more substrates in memory than templates in disk and stateful VMs can be created from the same or different substrates and deployed on to the same or different physical hosts in a cluster without causing any configuration conflicts. Experimental results show that general purpose VMs or a VM cluster for parallel computing can be deployed in a few seconds. We demonstrate the usage of VM substrates in a mobile gaming application.

Live migration of virtual machines (VM) across physical hosts provides a significant new benefit for administrators of data centers and clusters. Previous memory-to-memory approaches demonstrate the effectiveness of live VM migration in local area networks (LAN), but they would cause a long period of downtime in a wide area network (WAN) environment. This paper describes the design and implementation of a novel approach, namely, CR/TR-Motion, which adopts checkpointing/recovery and trace/replay technologies to provide fast, transparent VM migration for both LAN and WAN environments. With execution trace logged on the source host, a synchronization algorithm is performed to orchestrate the running source and target VMs until they reach a consistent state. CR/TR-Motion can greatly reduce the migration downtime and network bandwidth consumption. Experimental results show that the approach can drastically reduce migration overheads compared with memory-to-memory approach in a LAN: up to 72.4 percent on application observed downtime, up to 31.5 percent on total migration time, and up to 95.9 percent on the data to synchronize the VM state. The application performance overhead due to migration is kept within 8.54 percent on average. The results also show that for a variety of workloads migrated across WANs, the migration downtime is less than 300 milliseconds.

Understanding server capacity is crucial to system capacity planning, configuration, and QoS-aware resource management. Conventional stress testing approaches measure server capacity offline in terms of application-level performance metrics like response time and throughput. They are limited in measurement accuracy and timeliness. In a multitier website, resource bottleneck often shifts between tiers as client access pattern changes. This makes the problem of online capacity measurement even more challenge. This paper presents an online measurement approach based on low-level hardware performance metrics such as instructions execution rate and cache access behavior. Such metrics together define a system internal running state. The measurement approach uses machine learning techniques to infer application-level performance at each tier from a set of selected hardware performance counters. A coordinated predictor is induced over individual tier-wide models to make global system performance prediction and identify the bottleneck when the system becomes overloaded. Experiments were conducted on a two-tier Tomcat/MySQL-configured website using TPC-W benchmarks. Experimental results demonstrated that this approach was able to achieve an overload prediction accuracy of higher than 90 percent for a priori known input traffic mix and over 85 percent accuracy even for traffic causing frequent bottleneck shifting. It costs less than 0.5 percent runtime overhead for data collection and no more than 50 ms for each online decision making.