3D LiDAR scene completion from point clouds is a fundamental component of perception systems in autonomous vehicles. Previous methods have predominantly employed diffusion models for high‑fidelity reconstruction. However, their multi-step iterative sampling incurs significant computational overhead, limiting its real-time applicability. To address this, we propose LiNeXt: a lightweight, non‐diffusion network optimized for rapid and accurate point cloud completion. Specifically, LiNeXt first applies the Noise‑to‑Coarse (N2C) Module to denoise the input noisy point cloud in a single pass, thereby obviating the multi‑step iterative sampling of diffusion‑based methods. The Refine Module then takes the coarse point cloud and its intermediate features from the N2C Module to perform more precise refinement, further enhancing structural completeness. Furthermore, we observe that LiDAR point clouds exhibit a distance-dependent spatial distribution, being densely sampled at proximal ranges and sparsely sampled at distal ranges. Accordingly, we propose the Distance‑aware Selected Repeat strategy to generate a more uniformly distributed noisy point cloud. On the SemanticKITTI dataset, LiNeXt achieves a 199.8 times speedup in inference, reduces Chamfer Distance by 50.7 percent, and uses only 6.1 percent of the parameters compared with LiDiff. These results demonstrate the superior efficiency and effectiveness of LiNeXt for real-time scene completion.

Recently, Test-Time Adaptation (TTA) has gained increasing attention in medical imaging due to its ability to improve model generalization under domain shifts without retraining. In particular, directly applying a well-trained model across various medical centers faces significant performance degradation caused by variations in equipment, operators, imaging conditions, and scanning skill levels of sonographers. Existing TTA methods either rely on parameter adaptation that increases computational cost or apply simple prediction fusion that ignores anatomical structure knowledge. To address these limitations, we propose a novel backward-free Topology-aware TTA framework named T^3 that integrates Structural Perception Modeling (SPM) and Box Regression Adaptation (BRA). SPM is implemented through an organ space heatmap generated via Gaussian kernel superposition. This heatmap encodes anatomical topology without requiring additional training or source data. BRA further improves localization and classification by fusing detection outputs based on the contribution of detected results to anatomically meaningful peak points from the heatmaps. Extensive experiments were conducted across six cross-domain scenarios, and the results demonstrate that our method achieves state-of-the-art cross-domain detection performance while maintaining high efficiency, offering a practical and robust solution for real-world medical diagnostic applications.

Fetal ultrasound screening is a uniquely complex diagnostic task involving the simultaneous assessment of multiple fetal organs—each with its own anatomical and clinical context—within a single examination. Automating report generation for such cases poses a significant challenge: unlike existing methods that focus on single-organ radiology tasks (e.g., chest X-rays), fetal ultrasound requires reasoning over a structured, multiple-to-multiple setting, i.e., multi-organ images corresponding to a multi-section report. In this paper, we introduce FetusR, the first large-scale dataset for multi-organ fetal ultrasound reporting, containing 15,594 real-world cases with rich organ-wise annotations. To address the intrinsic image-report alignment, we propose Organ-Aware Routing Mixture-of-Retrieval Augmented Generation (ORM-RAG) inspired by the Mixture-of-Experts paradigm. Our method decomposes the complex alignment problem into multiple one-to-one sub-retrieval tasks. Specifically, ORM-RAG integrates (1) an organ-aware mixture-of-retrieval module that partitions the retrieval space into organ-specific corpora for independent retrieval, and (2) a dynamic routing mechanism that selectively aggregates high-confidence organ-specific reports while filtering uncertain ones. Extensive experiments demonstrate that our method significantly outperforms state-of-the-art baselines across both textual similarity and clinical accuracy metrics. Our work opens a new direction for long-form, structured report generation in real-world, multi-organ medical imaging scenarios.

Echocardiography and vascular ultrasound are essential for comprehensive cardiovascular assessment, yet manual evaluation and writing reports are labor-intensive, time-consuming, and require expertise from both cardiology and vascular surgery departments. Current automated report generation systems mainly focus on X-ray or CT, often neglecting echocardiographic modalities and critical quantitative parameters like aortic diameter and main pulmonary artery diameter, limiting their clinical utility. Moreover, the interdependence between cardiac and peripheral vascular health necessitates cross-departmental insights, which existing methods fail to incorporate. To address these limitations, we first propose the vision-language framework named the Echo-Cardiac-Vascular (ECV), for joint cardiac and vascular ultrasound report generation and parameter measurements. ECV introduces a Mixture-of-Experts vision encoder tailored for distinct ultrasound subtypes, a structured parameter measurement module for accurate quantification, and task-specific decoders that generate interpretable, multimodal diagnostic reports. Our framework, trained on 10K+ paired records, achieves high accuracy, improving diagnostic efficiency, consistency, and cross-disciplinary clinical applicability.

Haoliang Huang, Zihuang Cai, Zhuo Tang, Yifan Liu, Chen Tian, Kenli Li, Changjian Chen. Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). 2026.

Pruning is essential for the efficient deployment of Large Language Models (LLMs); however, it causes severe performance degradation due to the structural distortion induced by sparsity.Existing recovery strategies, such as LoRA, predominantly employ global finetuning, often overlooking the mechanistic root of this degradation: the layer-wise accumulation and amplification of local errors.To address this limitation, we propose LaCo (Layerwise Compensation), a framework that reorients the recovery paradigm from global adaptation to hierarchical representation alignment.By sequentially optimizing each layer to reconstruct the model's hidden states, LaCo effectively intercepts the error propagation chain at its source.Extensive experiments demonstrate that LaCo surpasses parameter-efficient baselines in both perplexity reduction and zeroshot reasoning.Notably, it reduces recoverytime memory usage to approximately 1/7 of the baseline and requires only 2,048 unlabeled samples to match a LoRA model trained on 50k examples-achieving a ∼ 25× improvement in data efficiency.

Sparse-Dense Matrix Multiplication (SpMM) on GPUs has gained significant attention because of its importance in modern applications and the increasing computing power of GPUs in the last decade. Previous SpMM studies have focused on the importance of storage format and load balance for the overall performance of SpMM on GPUs. However, very little attention has been paid to the efficacy of coalesced memory access in improving the efficiency of data loading, which incurs a notable overhead that amounts to an average of more than 32% of the overall performance, according to our experimental observation. Existing state-of-the-art (SOTA) solutions fail to adequately support coalesced memory access of both sparse and dense matrices between the global memory and threads on GPUs. In this paper, we propose an efficient algorithm called Swift.11Swift is available at https://github.com/MinttHu/Swift.git that speeds up the loading of both sparse and dense matrices of SpMM on modern GPUs. Leveraging coalesced memory access, Swift achieves high loading efficiency by sorting both the columns of the sparse matrix and elements of the dense matrix based on the number of non-zero elements and balancing the load by handling the regular and irregular parts differently and judiciously. Swift takes the Compressed Sparse Column format as an implementation case study to prove the concept and gain insights. We conduct a comprehensive comparison of Swift with four SOTA solutions: ASpT, cuSPARSE, RoDe, and Sputnik, using the full SuiteSparse Matrix Collection as the workload. The experimental results on RTX 4080s, RTX 3090Ti, A100, and V100 demonstrate that our method outperforms the baselines significantly.

Deep learning has made remarkable progress in medical image analysis, yet its black-box nature still limits interpretability and clinician trust. Concept-based modeling offers a promising direction for explainable AI by integrating human-understandable concepts. However, existing approaches typically rely on global concept annotations and infer diagnosis based solely on the presence or absence of individual concepts. This oversimplified paradigm ignores the rich relationships among concepts and their causal influence on disease outcomes. To overcome these limitations, we propose the Concept Relationship Embedding Model (CREM) for interpretable medical diagnosis. CREM mirrors coarse-to-fine clinical reasoning by first extracting fine-grained subregional concepts, then explicitly encoding their relationships as a concept interaction graph, and finally performing causal inference between concepts and diagnoses to enable reliable and transparent diagnostic predictions. We evaluate CREM on four public medical imaging benchmarks, where it achieves state-of-the-art performance on both concept recognition and disease classification tasks, while exhibiting improved robustness, label efficiency, and interpretability. Furthermore, we deploy CREM as an interactive web-based demo that allows clinicians to visualize concept activations, trace diagnostic reasoning paths, and iteratively refine concept cues, facilitating effective human-in-the-loop decision-making.

Existing tile size selection approaches are tightly coupled with compiler transformation pipelines, often leading to inaccurate modeling of cache behavior and limited effectiveness for non-rectangular tile shapes. This article presents TileMind , a decoupled analytical model that combines compile-time and runtime information for tile size selection in affine programs. It introduces a transformation-aware pre-tiling step that enables the decoupled selector to remain consistent with compiler transformations while extracting compile-time metadata. The extracted metadata is then combined with profiled runtime characteristics to construct a richer yet tractable feasible domain, within which a nonlinear objective for tile size selection is formulated. This objective is subsequently transformed into a binary product linearization problem, with its nonlinear constraints also linearized for efficient optimization. Finally, an intra-tile optimization aligns computation with data layout to enhance data reuse within tiles. Across two multi-core Intel CPUs, TileMind achieves 1.49× (sequential) and 1.33× (parallel) mean speedups on twenty PolyBench kernels, and 2.08–3.54× speedups on three deep learning workloads over the state-of-the-art analytical model Pluto-tss . Compared with TVM’s latest autotuner MetaSchedule, TileMind delivers 1.35–1.46× mean speedups while reducing tuning overhead by 2–4 orders of magnitude. While demonstrating effectiveness on selecting tile sizes for non-rectangular tile shapes and compatibility with PPCG, Pluto, and TVM, we further provide proof-of-concept results on GPUs, illustrating the potential portability of TileMind across architectures.

The Conjugate Gradient method is a prevalent iterative approach for solving sparse linear systems$Ax = b$, whereAis a symmetric positive definite matrix. In this scenario, the Factorized Sparse Approximate Inverse (FSAI) preconditioner is commonly used. The FSAI approximates$A^{-1}$by the matrix productGTG, whereGis a lower triangular matrix. The numerical properties of FSAI mainly depend on the sparsity pattern ofG. This work presents an approach to generate FSAI sparsity patterns tailored to SIMD architectures, which play a pivotal role in current high-performance computing systems. First, we implement FSAI-S, an FSAI preconditioner based on the SELL-C-σ format for SIMD architectures. To further improve iterative performance, we propose a padding-aware preconditioner, FSAI-P, which leverages the zero padding inherent in the SELL-C-σ to extend the sparsity pattern with minimal computational and storage overhead. We evaluate our approach on three SIMD architectures: a Skylake processor implementing the AVX-512 ISA, a RISC-V processor supporting the RISC-V “V” Vector ISA, and an Nvidia A100 GPU. Experimental results show that FSAI-P provides performance gains across all evaluated platforms. Specifically, average speedups are 12.51% (vs. FSAI) and 11.66% (vs. FSAI-cache) on Skylake, 57.79% (vs. FSAI) on RISC-V, and 18.33% (vs. FSAI) on the A100 GPU.

Spin Transfer Torque-Magnetic Random Access Memory (STT-MRAM), as a new non-volatile memory technology with lower leakage power and higher density, is widely considered to be a new generation of memory technology that may replace SRAM in the cache. STT-MRAM is divided into Single-Level Cell (SLC) STT-MRAM and Multi-Level Cell (MLC) STT-MRAM. Compared with SLC STT-MRAM, MLC STT-MRAM has further improved its storage density. However, MLC STT-MRAM has a high energy consumption and write latency due to its unique two-step state transitions (TTs) issue. State-of-the-art approaches mitigate this issue by eliminating TTs with expansion coding methods. Unfortunately, they focus more on eliminating TTs and have limited improvement in reducing energy consumption. To this end, we propose a new scheme, AEIS, which further reduces energy consumption while eliminating TTs. Our work begins with exploring the general rules of (M,N)-based expansion coding methods that eliminate TTs. Based on the discovered rules, the minimum energy coding method is found. To further improve energy efficiency, we segment the cache lines according to the data pattern. We only apply the expansion coding to those flipping segments to reduce the expansion coding overhead. The evaluation results show that AEIS can eliminate TTs in MLC STT-MRAM, reduce energy consumption by 28.5%, and increase the lifetime by 24.8%, while the total number of bits used for the cache only increases by 5.7%.

As integrated circuit design continues to scale, three-dimensional integrated circuit (3D-IC) technology has emerged as a promising solution to extend Moore’s Law. Among various physical design stages, placement plays a pivotal role, directly influencing key power, performance, and area (PPA) metrics. In research that does not consider specific 3DIC process technologies, the state-of-the-art ePlace-3D introduces 3D electrostatic field density modeling, enabling nonlinear placement algorithms to truly address 3D placement problems. However, its complex flow results in low runtime efficiency. In this work, we propose LightPlace, a lightweight and effective 3D placement framework. This framework incorporates Virtual Macro Insertion approach and Macro Connection Reconstruction method, which streamline the macro placement phase and allocate space for subsequent standard cell placement. These innovations enhance runtime efficiency and optimize wirelength. Experimental results on the large-scale Modern Mixed-Size (MMS) benchmarks demonstrate that, LightPlace reduces total wirelength by an average of 10% and achieves up to 4× faster runtime compared to the runtime reported for ePlace-3D.

Single-domain generalized object detection aims to enhance a model’s generalization to multiple unseen target domains using only data from a single source domain during training. This is a practical yet challenging scenario, as it requires the model to address domain shift without incorporating target domain data into the training process. In this paper, we propose a novel phrase-grounding-based style transfer (PGST) approach for the task. Specifically, we first define textual prompts to describe objects for potential unseen target domains. Then, we leverage the grounded language-image pre-training (GLIP) model to capture the styles of these target domains and perform style transfer from the source to the target domains. The style-transferred visual features from the source domain are semantically rich and closely approximate those of their hypothetical counterparts in the target domain. Finally, we employ these style-transferred visual features to fine-tune GLIP. By introducing these imaginary counterparts, the detector can be effectively generalized to unseen target domains using only a single source domain during training. Our method significantly improves mean average precision (mAP), with an average increase of 8.8% across five diverse weather-driving benchmarks. Notably, our approach outperforms or matches the performance of domain-adaptive object detection methods, which require target domain data for training, in several challenging scenarios.

Early disease diagnosis is critical for timely clinical intervention and treatment. Intelligent models have shown significant potential in addressing the challenges of misdiagnosis, especially given the shortage of experienced experts. However, there exists complexity of ultrasound image information, and subtle differences between positive and negative samples, combined with limited disease data, pose challenges in feature extraction, class imbalance, and the lack of representative positive class prototypes. To this end, we propose a large vision model distillation framework with fuzzy perception to improve rare disease diagnosis. Specifically, we first fine-tune the pre-trained Medical Segment Anything Model (MedSAM) to adapt it to the target domain. Through image augmentation and latent feature relationship distillation, we enhance feature extraction robustness and reduce inter-class ambiguity, which helps mitigate the impact of class imbalance on the lightweight student model. Second, as imaging style differences in ultrasound images are often more pronounced than subtle variations between positive and negative samples, we construct image style subsets and introduce a fuzzy style matching strategy to perceive these differences. Finally, we combine features from both the differential perception and semantic enhancement branches to strengthen disease classification. Extensive experiments on an internal fetal spina bifida dataset and three widely used imbalanced benign-malignant medical datasets demonstrate the effectiveness of the proposed method.

Matrix operations (MOPs) are essential for various computational tasks, particularly in deep learning models, which have grown increasingly complex. As these models expand, their demand for computational resources increases significantly, making deployment on resource-limited hardware platforms, such as field-programmable gate arrays, more challenging, especially in balancing resource allocation and computation latency. Reuse-control techniques have been employed to optimize resource allocation by enabling multiple operations to share the same computational units, such as digital signal processors. Yet, this approach presents a tradeoff between resource utilization and latency. In this study, we tackle this challenge by thoroughly analyzing existing reuse-control mechanisms and introducing a novel integer linear programming (ILP)-based strategy. Our experimental results demonstrate that the proposed approach not only improves resource utilization for large-scale MOPs but also significantly reduces latency compared to existing methods. In the best case, on the ResNet model, our ILP-based method achieves up to$2.21\times$lower latency and$2.14\times$lower energy consumption per inference, demonstrating significantly improved performance and energy efficiency. In addition, our work provides a new optimization perspective for hardware design based on high-level synthesis.

Closed spina bifida is a high-incidence developmental disorder among rare fetal diseases. Its signs in ultrasound imaging are subtle, making it prone to misdiagnosis and heavily reliant on sonographers' experience. Therefore, we propose a novel semantic enhancement framework incorporating projected attention for the automated screening of closed spina bifida through precise landmark detection. In this method, we utilize a multi-granularity deep supervision and voting mechanism to generate point-specific features and reconstruct saliency maps for each landmark, effectively reducing interference from homogeneous high-echogenic noise in ultrasound images while preserving rich semantic information. Additionally, a coordinate attention projection module is designed to convert the 2D landmark probability maps into one-dimensional vectors, ensuring low computational complexity along with precise coordinate regression. The clinical application potential of this intelligent system is significant, as it facilitates automated fetal spine counting and anatomical measurement, enabling early warnings of diseases based on identified anomalies. Extensive experiments comparing our method with advanced baselines on an in-house dataset and two public datasets demonstrate its clear advantage in computational complexity and accuracy.

Fetal anatomical structure segmentation in ultrasound images is essential for biometric measurement and disease diagnosis. However, current methods focus on a specific plane or a few structures, whereas obstetricians diagnose by considering multiple structures from different planes. In addition, existing methods struggle with segmenting fuzzy regions, which leads to performance degradation. We propose a real-time segmentation method called Class-aware Multi-structure Instance Segmentation (CMIS), designed to segment 19 key structures in 3 fetal brain planes to support brain-disease diagnosis. We extract instance information and generate class-aware attention for each class instead of dense instances to save computing resources and provide more informative details. Then we implement cross-layer and multi-scale fusion to obtain detailed prototypes. Finally, we fuse global attention with local prototypes cropped by boxes to generate masks and randomly perturb the boxes during training to enhance robustness. Moreover, we propose a new fuzzy region-based constraint loss to address the challenge of structures with varying scales and fuzzy boundaries. Extensive experiments on a fetal brain dataset demonstrate that CMIS outperforms 13 competing baselines, with an mDice of 83.41$\pm$0.03% at 37 FPS. CMIS also excels in external experiments on a fetal heart ultrasound dataset, achieving a mDice of 85.73$\pm$0.02% . These results demonstrate the effectiveness of CMIS in segmenting complex anatomical structures in ultrasound and its potential for real-time clinical applications. CMIS is limited to 2D normal standard planes ($\geq$19 weeks). Thus, its generalization to abnormal cases and broader datasets remains to be investigated.

Knowledge Tracing (KT) aims to predict learners' future responses by modeling historical learning sequences. However, existing works often overlook the severe disruptions in temporal and interaction continuity arising from large time intervals between interactions, which leads to significant prediction errors. Therefore, we identify and validate thisbreakpointeffect through empirical studies, and try to address it via data augmentation (DA). Existing DA methods usually emphasize sequence diversity or similarity while neglecting the breakpoint effect in KT. Hence, we strive to enhance knowledge tracing via breakpoint-aware sequence augmentation and propose Collaborative Learning Path Augmentation (CLPA), a systematic three-phase method to enhance models' representational capacity and alleviate the breakpoint effects. First, identifying boundary interaction pairs that exhibit the temporal gap and response inconsistency. Second, leveraging cross-sequence collaborative information to infer functional pseudo-paths between identified boundaries. Third, reconstructing and inserting pseudo-learning paths into target sequences to restore sequence continuity; moreover, two refinements of the sampling quantity constraint and the distance-based sampling strategy are proposed to ensure augmentation quality. Comprehensive experiments on four real-world datasets validate the effectiveness of CLPA, which significantly enhances the target KT model's representational capacity and alleviates the breakpoint effects.

Graph pattern queries (GPQ) over RDF graphs extend basic graph patterns to support variable-length paths (VLP), thereby enabling complex knowledge retrieval and navigation. Generally, variable-length paths describe the reachability between two vertices via a given property within a specified range. With the increasing scale of RDF graphs, it is necessary to design a partitioning method to achieve efficient distributed queries. Although many partitioning strategies have been proposed for large RDF graphs, most existing methods result in numerous inter-partition joins when processing GPQs, which impacts query performance. In this paper, we formulate a new partitioning problem, MaxLocJoin, aims to minimize inter-partition joins during distributed GPQ processing. For MaxLocJoin, we propose a partitioning framework (PIP) based on property-induced subgraphs, which consist of edges with a specific set of properties. The framework first finds a locally joinable property set using a cost-driven algorithm, LJPS, where the cost depends on the sizes of weakly connected components within its property-induced subgraphs. Subsequently, the graph is partitioned according to the weakly connected components. The framework can achieve two key objectives: first, it enables complete local processing of all variable-length path queries (eliminating inter-partition joins); second, it can minimize the number of inter-partition joins required for traditional graph pattern queries. Moreover, we identify two types of independently executable queries (IEQ): the locally joinable IEQ and the single-property IEQ. After that, a query decomposition algorithm is designed to transform all GPQ into one of them for independent execution in distributed environments. In experiments, we implement two prototype systems based on Jena and Virtuoso, and evaluate them over both real and synthetic RDF graphs. The results show that MaxLocJoin achieves performance improvements from 2.8x to 10.7x over existing methods.

Spatial crowdsourcing (SC) is becoming increasingly popular recently. As a critical issue in SC, task assignment currently faces challenges due to the imbalanced spatiotemporal distribution of tasks. Hence, many related studies and applications focusing on cross-platform task allocation in SC have emerged. Existing work primarily focuses on the maximization of total revenue for inner platform in cross task assignment. In this work, we formulate a SC problem called Cross Dynamic Task Assignment (CDTA) to maximize the overall utility and propose improved solutions aiming at creating a win-win situation for inner platform, task requesters, and outer workers. We first design a hybrid batch processing framework and a novel cross-platform incentive mechanism. Then, with the purpose of allocating tasks to both inner and outer workers, we present a KM-based algorithm that gets the accurate assignment result in each batch and a density-aware greedy algorithm with high efficiency. To maximize the revenue of inner platform and outer workers simultaneously, we model the competition among outer workers as a potential game that is shown to have at least one pure Nash equilibrium and develop a game-theoretic method. Additionally, a simulated annealing-based improved algorithm is proposed to avoid falling into local optima. Last but not least, since random thresholds lead to unstable results when picking tasks that are preferentially assigned to inner workers, we devise an adaptive threshold selection algorithm based on multi-armed bandit to further improve the overall utility. Extensive experiments demonstrate the effectiveness and efficiency of our proposed algorithms on both real and synthetic datasets.

Graph-structured data appears in domains such as molecular analysis, social networks, and program optimization, where graphs often exhibit implicit heterogeneity, as nodes may look homogeneous in type yet differ significantly in semantics or functionality. Graph Neural Networks (GNNs), while powerful on homophilic graphs, tend to degrade in such settings due to polarity confusion, over-smoothing, and inefficiency caused by dense propagation. We propose a polarity-aware framework for graph classification that addresses these challenges through adaptive directional sparse aggregation. The framework introduces a polarity-aware propagation mechanism that adaptively reinforces or inverts neighbor signals, mitigating contamination under heterophily. A polarity-guided sparse aggregation operator further alleviates over-smoothing, improves scalability by constraining redundant connections, and condenses information flow into more effective representations, while maintaining unbiased estimation with controlled variance. We provide theoretical analyses that characterize the computational complexity, stability properties, and expressive behavior of signed directional aggregation, offering theoretical insights into its computational, stability, and expressive properties. Extensive experiments on molecular and social graph benchmarks with implicit heterophily demonstrate consistent improvements in graph classification accuracy and efficiency. Our method achieves a 2.36% improvement when compared with the strongest baseline on each dataset. In addition, it improves accuracy by 4.53% on average on program optimization strategy recognition tasks, reaching 80.12% overall.

In-memory dynamic graph processing faces three critical challenges: limited DRAM capacity, inefficient concurrent update/query handling, and vulnerability to crashes. Traditional segment-level systems struggle with write amplification on emerging Storage-Class Memory (SCM), while existing persistent-memory systems suffer from coarse-grained synchronization and high recovery overhead. This study presents the Storage-Class Memory Dynamic Graph (SMDG) processing framework, an architecture-level redesign centered on the block as the atomic unit across storage, concurrency, and recovery. The system addresses these challenges through three key innovations. First, a block-granular storage design organizes adjacency data at fixed-size block granularity on heterogeneous DRAM-SCM architecture, employing buffered batched writes to significantly reduce write amplification while preserving logarithmic update complexity. Second, block-level multi-version concurrency control maintains timestamped block versions under per-vertex read-write synchronization to provide task-ordered snapshot visibility for concurrent queries without copying entire vertices or pages. Third, a block-granular crash recovery protocol with decentralized per-vertex logs enables independent parallel reconstruction, ensuring application-level semantic consistency while achieving substantially faster recovery than sequential approaches. Experimental results validate that this unified block-granular design improves update efficiency, sustains mixed update-query workloads with controlled memory overhead, and accelerates crash recovery compared with prior dynamic graph systems.

Federated learning (FL) enables collaborative training of a global model while preserving participants' local data privacy, making it ideal for data-sensitive fields like Industrial Internet of Things (IIoT), finance, and healthcare. However, Non-IID data among participants and the presence of malicious participants pose significant challenges to the model's performance and convergence. The global model is difficult to achieve consistent performance across all participants. Therefore, this paper proposes personalized and robust federated learning (PRFL) to handle non-independently and identically distributed (Non-IID) data with malicious participants. First, to enhance the robustness and convergence, a model similarity-based division mechanism is employed. It groups participants with similar data and removes both independent and colluding malicious participants. Second, we propose a three-stage knowledge sharing personalized federated learning framework. Each participant undergoes inner-loop knowledge sharing, outer-loop knowledge sharing, and personalized knowledge distillation, incorporating performance-driven dynamic weighted sharing mechanism. Moreover, extensive experiments demonstrate that PRFL outper forms other advanced personalized federated learning methods across various benchmark datasets, particularly in scenarios with Non-IID data and malicious participants.

Backward-compatible training enables the deployment of advanced models without requiring updates to old gallery databases. However, existing methods, including old-prototype-based (i.e., those relying on prototypes from the old model) and instance-based approaches, often overlook the impact of the old model's quality. High-quality old models exhibit compact intra-class feature distributions, which facilitate effective alignment between old and new models across various methods. In contrast, low-quality old models produce dispersed features, making it difficult for old-prototype-based methods to extract sufficient information. Additionally, instance-based methods are overly restrictive, limiting the flexibility of new models. In this work, we propose SA-BCT, an extremely simple yet effective backward-compatible training method that offers a unified framework for accommodating old models of varying quality. SA-BCT employs a single loss function applied to both old and new features, self-adaptively adjusting the constraint space for new features based on the distribution of old features. Extensive experiments in diverse settings demonstrate the effectiveness of SA-BCT. Code is available athttps://github.com/yuleung/SA-BCT.

Camera-based Semantic Scene Completion (SSC) aims to infer the geometric structure and semantic information in the entire 3D scene from limited 2D images. However, due to the lack of geometric information in the image, existing methods tend to generate fuzzy completion and incorrect semantic boundaries. In this paper, we propose cross-modal knowledge distillation to address this issue, namely PI-Net, which guides the camera-based model to learn accurate 3D geometry to compensate for spatial surroundings information during training. Specifically, we propose a point cloud occupancy prediction model as the teacher, leveraging its output for strong depth supervision signals and spatial voxel information to enhance the student model. To facilitate effective distillation, we design depth guidance distillation to improve geometric predictions, and spatial guidance distillation to assist the student model in better capturing the structural information of the surrounding environment. Finally, prediction domain distillation is incorporated to facilitate holistic learning from point cloud to image. Experimental results demonstrate that PI-Net outperforms state-of-the-art camera-based methods on challenging benchmarks—SemanticKITTI and SSCBench-KITTI-360.

The criminal court view generation (CCVG) task aims to produce succinct and coherent summaries of fact descriptions, providing interpretable opinions for verdicts. Traditional text generation evaluation metrics, such as ROUGE, BLEU, and BERTSCORE, are extensively employed for this task and measure performance by averaging the assessment scores of all samples within the test set. However, these sample-averaged metrics encounter two primary dilemmas: 1) they fail to fairly assess overall evaluation scores across different case types and 2) they overlook the measurement of the degree of performance imbalance between case types. To fill this research gap, we propose two novel case-type-oriented evaluation metrics: Case-type-oriented Text Generation (CTG) and Case-type-oriented Imbalance Performance (CIP). First, CTG mitigates the unfair assessment among different case types by assigning equal weight to each type. Second, CIP evaluates performance imbalance by measuring the distance between the performance of each case type and the overall performance. We provide three theorems to elucidate the properties of CIP, demonstrating that CIP can effectively identify the extent to which a CCVG model achieves balanced generation performance across different case types. Furthermore, we propose an embarrassingly simple and effective charge-guided encoder-decoder (CGED) framework to enhance performance fairly across different case types in encoder-decoder pretrained language models (PLMs). Code is available at https://yuquanle.github.io/Case-type-oriented-metrics-homepage/.

Sparse tensors are prevalent in real-world applications, often characterized by their large-scale, high-order, and high dimensional nature. Directly handling raw tensors is impractical due to the significant memory and computational overhead involved. The current mainstream approach involves compressing or decomposing the original tensor. One popular tensor decomposition algorithm is the Tucker decomposition. However, existing state-of-the-art algorithms for large-scale Tucker decomposition typically relax the original optimization problem into multiple convex optimization problems to ensure polynomial convergence. Unfortunately, these algorithms tend to converge slowly. In contrast, tensor decomposition exhibits a simple optimization landscape, making local search algorithms capable of converging to a global (approximate) optimum much faster. In this paper, we propose the FastTuckerPlus algorithm, which decomposes the original optimization problem into two non-convex optimization problems and solves them alternately using the Stochastic Gradient Descent method. Furthermore, we introduce cuFastTuckerPlusTC, a fine-grained parallel algorithm designed for GPU platforms, leveraging the performance of tensor cores. This algorithm minimizes memory access overhead and computational costs, surpassing the state-of-the-art algorithms. Our experimental results demonstrate that the proposed method achieves a 2× to 8× improvement in convergence speed and a 3× to 5× improvement in per-iteration execution speed compared with state-of-the-art algorithms.

Deep neural networks (DNNs) dominate workloads on cloud and edge platforms. Meanwhile, the hardware platform towards the heterogeneous system with various accelerators. By mapping layers to their different preferred accelerators, the computation cost of each layer can be reduced. While mapping these layers on the same accelerator can reduce the inter-accelerator communication cost. These two costs are often competing and difficult to optimize simultaneously. Therefore, the core challenge in achieving efficient execution of DNN workloads on heterogeneous systems is: how to map layers to achieve the best trade-off between computation and communication costs. Existing works group layers into blocks and perform blockwise mapping to reduce inter-layer communication within blocks. However, when grouping layers, they typically rely on modelagnostic rules, which fail to hide critical inter-layer communication within blocks for diverse DNNs. Moreover, after block mapping, the lack of intra-block resource allocation further increases computation cost of block. In this paper, we proposeGHCoM, a novel block-wise mapping framework for exploring the effective cost trade-offs.GHCoMemploys an adaptive grouping strategy to guide layer grouping based on the topology of DNNs and dynamically adjust the grouping according to the trade-off target. Furthermore,GHCoMconsiders the fine-grained allocation of computation (i.e., processing elements) and communication (i.e., on-chip bandwidth) resources within each block to mitigate interlayer resource contention. To jointly optimize layer grouping, block-wise mapping and intra-block resource allocation,GHCoMleverages a two-level genetic algorithm (GA) with tailored encodings and operators that capture the interdependence across the entire design space. Experiments across various workloads and system configurations show thatGHCoMconsistently outperforms state-of-the-art baselines, achieving 1.08× to 4.79× speedup in execution latency and reducing energy consumption by 1.83% to 87.71%.

Graph-structured data has been widely applied in transportation, molecular, and e-commerce networks, etc. Graph Convolutional Network (GCN) has emerged as an efficient approach to processing non-Euclidean graph data. However, the varying sizes and sparsity of graph datasets, coupled with the dependency of the dataflow patterns in GCN computation on the graph data, have rendered the acceleration of GCN inference increasingly challenging. This paper proposes a GCN inference accelerator based on multi-dataflow and high bandwidth memory (HBM), named DAHBM-GCN. Firstly, we designed a computing engine that supports multiple dataflows, aggregation-first, and combination-first orders. Furthermore, an adaptive selector for the multi-dataflow computing engine based on the decision tree is proposed to select the optimal dataflow computing engine. Secondly, an efficient mapping of pseudo channels (PCs) for multi-channel HBM is devised to enhance bandwidth, effectively alleviating memory latency and bandwidth bottlenecks. Thirdly, a hybrid fixed-point quantization strategy for GCN is introduced, which reduces the GCN model's computation complexity and parameter count with almost no loss of accuracy. Finally, extensive performance evaluation experiments demonstrate that across various datasets, DAHBM-GCN achieved average speedups of 52.5–129.3× and 4.9–7.9× compared to PyG-GCN and DGL-GCN on CPU, respectively. Compared to the AWB-GCN, HyGCN, HLS-GCN, and GCNAX accelerators FPGA-based, DAHBM-GCN also exhibits average speedups of 1.21–2.21×, 1.25–1.98×, 1.65–2.68×, and 1.18–1.56× respectively, on various datasets. Additionally, DAHBM-GCN possesses the advantages of high flexibility and low energy consumption.

Federated Learning (FL) has emerged as a key framework to deliver AI services, recognized for its capability to construct global models while ensuring individual data. Nevertheless, FL heavily relies on a central server, which introduces significant challenges for participants to collaborate effectively and substantially limits the scalability of FL. Blockchain-based FL (BFL) offers a promising solution by replacing the central server with a decentralized blockchain system, thereby establishing a secure and trustworthy environment for FL. However, current BFL approaches face challenges in balancing high computational overhead, consistency, and security. In view of this, this paper introduces EBFL, an efficient blockchain framework for FL services. EBFL incorporates both asynchronous and synchronous advantages. A DAG-based (Directed Acyclic Graph) asynchronous computation enhances computational efficiency by mitigating delays caused by slow devices and reducing unnecessary waiting due to frequent synchronized consensus. Simultaneously, a periodic synchronized consensus mechanism is introduced during asynchronous training to ensure consistency, thereby improving security and model accuracy. Additionally, taking into account the unique characteristics of FL, we have designed a series of operations tailored for EBFL to further enhance the performance. Experimental results demonstrate that, compared to traditional synchronous BFL (TBFL) approaches, EBFL achieved a maximum speedup of up to 2.38× while retaining 92% of their accuracy. Subsequently, in-depth analytical experiments show that EBFL excels in both convergence speed and security, thereby confirming its potential to balance computational efficiency, consistency, and security.

Hybrid rice breeding crossbreeds different rice lines and cultivates the resulting hybrids in fields to select those with desirable agronomic traits, such as higher yields. Recently, genomic selection has emerged as an efficient way for hybrid rice breeding. It predicts the traits of hybrids based on their genes, which helps exclude many undesired hybrids, largely reducing the workload of field cultivation. However, due to the limited accuracy of genomic prediction models, breeders still need to combine their experience with the models to identify regulatory genes that control traits and select hybrids, which remains a time-consuming process. To ease this process, in this paper, we proposed a visual analysis method to facilitate interactive hybrid rice breeding. Regulatory gene identification and hybrid selection naturally ensemble a dual-analysis task. Therefore, we developed a parametric dual projection method with theoretical guarantees to facilitate interactive dual analysis. Based on this dual projection method, we further developed a gene visualization and a hybrid visualization to verify the identified regulatory genes and hybrids. The effectiveness of our method is demonstrated through the quantitative evaluation of the parametric dual projection method, identified regulatory genes and desired hybrids in the case study, and positive feedback from breeders.

High-quality remote sensing imagery plays a vital role in environmental monitoring and disaster management. However, transmitting these images is challenging due to the unstable signal-to-noise ratio (SNR) and bandwidth limitations encountered in remote communications. Semantic communication, particularly deep learning-based methods, offers a promising solution by jointly optimizing source and channel coding to achieve data compression and noise resilience. Nevertheless, existing methods struggle to cope with varying channel noise and bandwidth, leading to unsatisfactory image reconstruction quality. To address these challenges, we propose a satellite-ground compression and transmission system called Adaptive Residual Joint Source-Channel Coding (ARJSCC), which is based on Deep Joint Source-Channel Coding (DeepJSCC). The ARJSCC system compresses remote sensing images into semantic information and residuals to achieve low overhead transmission and high-quality reconstruction. ARJSCC utilizes an attention module to adjust the semantic preference of the model for different SNRs, and deploys a variance-based position mask module to flexibly vary the semantic length and further compress it. These designs enable ARJSCC to automatically adapt to varying noise and bandwidth conditions. Moreover, for the residual, we apply BPG to compress it to reduce the transmission cost and design the corresponding enhancement module to recover its details from the noise-affected compressed residual. We experimentally compare our ARJSCC with the recent DeepJSCC-based wireless image transmission models in low-resolution dataset and high-resolution remote dataset under multiple wireless channel environments. The experimental results show that ARJSCC can achieve high reconstruction quality exceeding 44dB, and outperform the competitors by 4-6db even under low SNR and bandwidth environments.

Source-free unsupervised domain adaptation aims to eliminate domain shifts when data from the source domain and annotation from the target domain are not available. The multi-object detection tasks in medical image analysis are constrained by patient privacy and extremely huge annotation consumption. Hence, Source-free UDA is considered a more practical approach for eliminating the domain gap. However, relevant research that explores this topic is a dearth. In this paper, we design an Anatomy-aware Alignment Teacher-Student learning method using topological consistency based on a mean-teacher framework for Source-free UDA in multiple medical object detection named AATS, including Unsupervised Structure Refinement (USR) and Graph-aware Morphology Alignment (GMA). To match the student and teacher at the low-level and visual features, we propose the USR via an unsupervised clustering algorithm to group organs in ultrasound images. Based on USR, we obtain a graph with organ relations on the teacher branch. While in the student branch, we acquire visual features to construct graphical space and optimize the model with graph propagation. Finally, to match the student and teacher, GMA is designed to align the teacher and student based on both topology and morphology information that is derived from prior medical knowledge. Four groups of adaptation experiments were conducted on available medical datasets, and the outcomes demonstrate that our approach not only achieves state-of-the-art performance but also provides substantial advantages over existing methods.

In medical image analysis, detecting multiple structures is crucial for evaluations and diagnosis but is often limited by the lack of high-quality annotations. Semi-supervised object detection emerges as a potent methodology to enhance model performance and generalization by leveraging a vast pool of unlabeled data alongside a minimal set of labeled data. A striking observation is that both unlabelled and labeled medical images contain a priori anatomical knowledge from human screening. In this work, we introduce a novel semi-supervised approach named Semi-akmm for mining and matching anatomical knowledge in ultrasound images. We develop an Adaptive Prior Knowledge Transfer (APKT) module to mine and explore the distribution and knowledge of potential proposal boxes by proposal proportion constraint. Furthermore, within a teacher-student learning framework, we put forward an Anatomical Structure Matching (ASM) module to facilitate co-learning consistent topological prior knowledge between the student and teacher models. To our knowledge, this marks the inception of an efficient semi-supervised medical multi-structure detection model. Our experiments across five publicly available ultrasound datasets demonstrate that Semi-akmm sets a new benchmark in performance with solid results that outperform existing methods.

Camera-based 3D semantic scene completion (SSC) provides dense geometric and semantic perception for autonomous driving. However, images provide limited information making the model susceptible to geometric ambiguity caused by occlusion and perspective distortion. Existing methods often lack explicit semantic modeling between objects, limiting their perception of 3D semantic context. To address these challenges, we propose a novel method VLScene: Vision-Language Guidance Distillation for Camera-based 3D Semantic Scene Completion. The key insight is to use the vision-language model to introduce high-level semantic priors to provide the object spatial context required for 3D scene understanding. Specifically, we design a vision-language guidance distillation process to enhance image features, which can effectively capture semantic knowledge from the surrounding environment and improve spatial context reasoning. In addition, we introduce a geometric-semantic sparse awareness mechanism to propagate geometric structures in the neighborhood and enhance semantic information through contextual sparse interactions. Experimental results demonstrate that VLScene achieves rank-1st performance on challenging benchmarks—SemanticKITTI and SSCBench-KITTI-360, yielding remarkably mIoU scores of 17.52 and 19.10, respectively.

Cardiovascular disease diagnosis via ultrasound screening relies on manually measured metrics and the experience level of human experts, which is time-consuming and may overlook subtle cross-anatomical pathological patterns. Recent vision-language models offer end-to-end diagnostic potential but lack mechanisms to handle heterogeneous multi-modal, multiview ultrasound data while preserving modality-specific semantics. To fill this gap, we propose an end-to-end framework called MMVL that directly fuses raw ultrasound sequences from diverse anatomical regions, bypassing intermediate measurements, and enabling direct diagnosis. We design lightweight adapters for domain-specific multi-modal feature fusion and refinement, a gating mechanism that dynamically reweights modality importance based on global context, and disease-aware prompt-guided classification. MMVL ensures robust performance across both common and rare conditions. The proposed multi-view, multimodal vision-language framework enables end-to-end cardiovascular disease diagnosis with a 10.9% accuracy gain, and opens a new avenue for automated and generalizable diagnostic solutions.

Molecular dynamics (MD) simulations are essential tools for investigating large-scale molecular systems, yet achieving high performance and scalability on CPU-based architectures remains challenging. In this study, we present a highly optimized framework based on DeepMD-kit for conducting 500 millionatom MD simulations on an ARMv8 SVE high-performance computing (HPC) system. Key optimizations include leveraging OpenMP for multi-threaded acceleration of DeepMD-kit and utilizing the ARMv8 SVE instruction set to optimize doubleprecision matrix multiplication in PyTorch. These enhancements enable single ARMv8 SVE 64-core processors to achieve 1.3x the training performance of NVIDIA V100 GPU, and two ARMv8 SVE 64-core processors to achieve 1.05x the inference performance of NVIDIA V100 GPU. Leveraging this optimized framework, we achieve large-scale MD simulations across 4,096 computing nodes.

Sparsity is widely prevalent in real-world applications, yet existing compiler optimizations and code generation techniques for sparse computations remain underdeveloped. Sparse matrix-matrix multiplication (SpMM) is a representative operator in sparse computations, whose performance is often limited by the design of sparse formats and the extent of hardware architecture optimization. Most existing solutions achieve highperformance SpMM through two approaches: (1) meticulously designed kernels and specialized sparse formats, which require extensive manual effort, or (2) tensor compilers that support code generation, though these typically offer limited support for sparse patterns, making it challenging to adapt to complex sparsity patterns in practical applications. This paper presents SpMMTC, an input-aware sparse tensor compiler. Given a sparse matrix as input, SpMMTC analyzes its non-zero distribution and generates a vectorized kernel optimized for SpMM on the specific matrix. We evaluated SpMMTC on various workloads. It achieves speedups of 1.21 x to 2.97 x over state-of-the-art methods such as TACO, TVM, and ASpT on different multi-core processors. It also provides a speedup of up to $\mathbf{1. 5 2 x}$ for sparse MobileNetV1 inference on the edge device.

The median (MED) is a crucial statistic for measuring the central tendency. However, exact MED computation remains costly, with even state-of-the-art (SOTA) algorithms failing to meet (near) real-time processing demands. While approximate MED algorithm has arisen as a promising candidate, existing approaches ignore the potential opportunity of spatiotemporal similarity within the application and fail to provide applicationspecific trade-offs between execution time and accuracy. Our goal is to design an enhanced approximate MED algorithm STREAM, which is capable of exploiting the spatiotemporal similarity to achieve bucket reuse and establish a tunable-grained bucket mechanism to meet the accuracy of application-specific requirements. Experimental results show that while maintaining nearly identical accuracy, STREAM outperforms the SOTA approximate methods DDSketch (up to $10 \times 4.7 \times$ on average) and KLL (up to $71.2 \times 10.1 \times$ on average).

Sparse Matrix-Vector Multiplication (SpMV) is an essential sparse operation in scientific computing and artificial intelligence. Efficiently adapting SpMV algorithms to diverse matrices and architectures requires a framework capable of accurately recognizing sparse patterns and selecting the optimal implementation. In this work, we introduce Sparsity-aware SpMV (SSpMV), a framework that integrates expert-designed features with multimodal representations to adaptively predict the best-performing algorithm and parameters. For this purpose, we design a multimodal neural network called MM-Adapter, to capture diverse modalities to represent the computational features of SpMV. Experimental results demonstrate that MMAdapter achieves the highest accuracy of $81.05 \%$, outperforming existing SpMV prediction models. Furthermore, SSpMV consistently delivers substantial performance improvements over state-of-the-art sparse libraries across various multi-core platforms.

High-Level Synthesis (HLS) offers various optimization directives that enable designers to flexibly adjust hardware microarchitecture. However, existing HLS performance prediction methods typically rely on control data flow graphs generated (CDFG) from original HLS C/C++, which struggle to capture the complex interactions between directives and the resulting hardware resource reuse issues. To address these issues, this paper proposes a post-implementation performance prediction method tailored for directive-optimized circuits design, which utilizes a Graph Builder to integrate directive optimization and resource reuse information into the graph representation. In addition, the performance prediction model, which integrates a TransformerConv-based graph neural network (GNN) and an aggregation pool module, effectively captures key features related to post-implementation performance. Experimental results show that our method can reduce the prediction error of critical path delay (CP), power, and resource utilization to 3.87% $\sim$ 8.08%, significantly outperforming existing state-of-the-art methods. It also demonstrates excellent generalization on unseen kernels, providing a more effective and accurate performance prediction tool for HLS.

Sparse general matrix-matrix multiplication (SpGEMM) is a core primitive for numerous scientific applications. Traditional hash-based approaches fail to strike a balance between reducing hash collisions and efficiently utilizing fast shared memory, which significantly undermines the performance of executing SpGEMM on GPUs. To address this issue, this paper introduces a novel accumulator design that achieves high shared memory utilization on modern GPUs. For the proposed high shared memory utilization algorithm, i.e., HSMU-SpGEMM1, we further optimize different symbolic stages. Our evaluations with four state-of-the-art hash-based SpGEMM libraries (Nsparse, spECK, OpSparse, and NVIDIA’s cuSPARSE) on three NVIDIA GPUs (Ampere, Ada Lovelace, Turing) demonstrate significant performance benefits from HSMU-SpGEMM.1HSMU-SpGEMM is available at https://github.com/wuminqaq/HSMUSpGEMM

Text-driven hair editing on 3D heads is a challenging problem in computer vision and graphics. In this paper, we propose TextHair3D, a NeRF-based text-driven 3D hair editing method that uses 3D perception to generate priors, edit hair attributes from user-provided text, and preserve facial features. TextHair3D uses the Contrastive Language-Image Pre-training (CLIP) model to encode textual conditions. To address the complexity and roughness of local editing, we design a combined conditional mapping module to map image and text conditions into latent space for learning generative priors. This enables high-quality, photo-realistic hair editing and 3D head reproduction. Extensive experiments show Tex-tHair3D’s superiority in visual realism and attribute accuracy.

Semi-supervised segmentation is gaining popularity in medical image analysis due to challenges in data acquisition and annotation. However, most methods focus on generating additional training pairs from unlabeled data through augmentation or perturbation for contrastive learning, often overlooking the unique characteristics and inherent priors of medical images. We identified two key anatomical priors in fetal cardiac ultrasound images: (1) anatomies have consistent shapes and locations due to standard views captured by sonographers from fixed angles; (2) category pixels are densely clustered, with each structure appearing only once per image. We propose AP-Net, which uses an anatomical prior generation module, a prior-feature fusion module, and a category-aware cropping strategy to effectively leverage these anatomical priors. Experiments on a real-world fetal cardiac ultrasound dataset show that AP-Net outperforms classical supervised and leading semi-supervised methods, with each component enhancing its performance.

With the rapid growth of multimodal and multi-view data, multi-view bipartite graph clustering (MVBGC) has emerged as a promising solution for large-scale tasks, which with linear complexity. However, most methods adhere to a unidirectional “encoding” design, where the bipartite graph is directly constructed from input data. Enlightened by the prevalent encoding-decoding in deep learning, this paper rethinks existing paradigms and proposes a novel “auto-encoding” MVBGC framework, named BGAE. Our model seamlessly integrates encoding, bipartite graph learning, and decoding modules within a self-supervised learning framework. The encoding module extracts a joint representation from input data, the bipartite graph learning module learns a discriminative bipartite graph in latent semantic space, and the decoding module reconstructs the input data by the structural information. Extensive experiments verify the superiority of our novel design, particularly highlighting the critical role of “decoding” learning. This work represents the first attempt to explore encoding-decoding design in MVBGC.

Structural Graph Clustering is a well-known problem that aims to identify clusters and distinguish between special roles, such as hub and outlier. However, SCAN, the fundamental structural clustering model, is designed for pairwise graphs and fails to capture the unique structural information inherent in hypergraphs when clustering hypergraphs. Motivated by this, we propose a new structural clustering model, HSCAN, specifically for hypergraphs. We further design an Order-Index to accelerate fetching the key information of the HSCAN and a Lightweight Similarity Bucket Index to reduce the index cost. Next, we present an index-based sequential query algorithm with high performance and a parallel query algorithm to process large hypergraphs faster. Additionally, we provide the algorithms for constructing Order-Index and Lightweight Similarity Bucket Index. Extensive experiments on both real-world and synthetic datasets show that HSCAN performs better than existing models, and the two index-based query algorithms are up to three orders of magnitude faster than the existing algorithm.

This paper addresses the pressing need for effective k-tips decomposition in dynamic bipartite graphs, a crucial aspect of real-time applications that analyze and mine binary relationship patterns. Recognizing the dynamic nature of these graphs, our study is the first to provide a solution for k-tips decomposition in such evolving environments. We introduce a pioneering projection-based algorithm, coupled with advanced incremental maintenance strategies for edge modifications, tailored specifically for dynamic graphs. This novel approach not only fills a significant gap in the analysis of dynamic bipartite graphs but also substantially enhances the accuracy and timeliness of data-driven decisions in critical areas like public health. Our contributions set a new benchmark in the field, paving the way for more nuanced and responsive analyses in various domains reliant on dynamic data interpretation.

Task assignment is a fundamental challenge in Spatial Crowdsourcing which aims to assign location-based tasks to workers under spatial-temporal constraints. Recently, some exciting research has introduced the preference of workers and tasks to improve assignment quality. However, they either primarily focus on the current preferences of both workers and tasks or only consider the unilateral prediction-based preference of workers, overlooking the impact of workers' interconnection and tasks' completed sequences. As a result, they gain suboptimal assignment results in most cases. Inspired by this, we propose a novel problem, named the Predictive Bi-preference Stable Match problem (PBSM), with the goal of maximizing the preferences of both workers and tasks by taking into account the social network of workers and task completion sequence. The PBSM problem is proven to be NP-hard. To tackle this challenging problem, we develop a GCN-enhanced Transformer-based Prediction and Bi-preference Stable Matching (GETBM) framework with two stages: the bi-preference prediction stage and the bilateral assignment stage. In the prediction stage, the Worker Preference Model (WPM) and Task Preference Model (TPM) models are presented to predict the worker-to-task (Worker2Task) and task-to-worker (Task2Worker) preference lists, respectively. Then, we design a bilateral preference-aware stable matching (BPM) algorithm and prove it can gain stable results. To generalize to multiple scenarios, three optimization strategies are devised based on spatial-temporal constraints and priority consideration to gain better assignment performance. Extensive experiments are conducted to prove the superiority of the GETBM framework on two real datasets.

As the network data in real life become multi-modal and multi-relational, multi-view attributed graphs have garnered significant attention. Numerous methods have achieved excellent performance in multi-view attributed graph clustering; however, they cannot efficiently handle incomplete attribute scenarios, which are prevalent in many real-life applications. Inspired by this, we investigate the problem of multi-view incomplete attributed graph clustering for the first time. In particular, the TOTF (Train Once Then Freeze) framework is designed to train missing-aware encoders that capture view-specific information while ignoring the impact of incomplete attributes, and then employs frozen encoders to uncover common information driven by clustering. After that, we propose a correlation strength-aware graph neural network on the basis of the inherent relationships among attributes to enhance accuracy. It is proven theoretically that traditional Generative Adversarial Networks (GANs) are unable to generate the unique real distribution. To address this issue, we further introduce the missing-position reminder mechanism into our intra-view adversarial games for better clustering results. Extensive experimental results demonstrate that our method achieves up to a 17% improvement in accuracy over the state-of-the-art methods. The source code is available at https://anonymous.4open.science/r/TOTF-main.

The performance of computer vision models in certain real-world applications, such as medical diagnosis, is often limited by the scarcity of available images. Expanding datasets using pre-trained generative models is an effective solution. However, due to the uncontrollable generation process and the ambiguity of natural language, noisy images may be generated. Re-weighting is an effective way to address this issue by assigning low weights to such noisy images. We first theoretically analyze three types of supervision for the generated images. Based on the theoretical analysis, we develop TriReWeight, a triplet-connection-based sample re-weighting method to enhance generative data augmentation. Theoretically, TriReWeight can be integrated with any generative data augmentation methods and never downgrade their performance. Moreover, its generalization approaches the optimal in the order O(√d ln (n)/n). Our experiments validate the correctness of the theoretical analysis and demonstrate that our method outperforms the existing SOTA methods by 7.9% on average over six natural image datasets and by 3.4% on average over three medical datasets. We also experimentally validate that our method can enhance the performance of different generative data augmentation methods.

Abstract Novel view synthesis for talking heads presents significant challenges due to the complex and diverse motion transformations involved. Conventional methods often resort to reliance on structure priors, like facial templates, to warp observed images into a canonical space conducive to rendering. However, the incorporation of such priors introduces a trade‐off‐while aiding in synthesis, they concurrently amplify model complexity, limiting generalizability to other deformable scenes. Departing from this paradigm, we introduce a pioneering solution: the motion‐conditioned neural radiance field, MoNeRF, designed to model talking heads through latent motion navigation. At the core of MoNeRF lies a novel approach utilizing a compact set of latent codes to represent orthogonal motion directions. This innovative strategy empowers MoNeRF to efficiently capture and depict intricate scene motion by linearly combining these latent codes. In an extended capability, MoNeRF facilitates motion control through latent code adjustments, supports view transfer based on reference videos, and seamlessly extends its applicability to model human bodies without necessitating structural modifications. Rigorous quantitative and qualitative experiments unequivocally demonstrate MoNeRF's superior performance compared to state‐of‐the‐art methods in talking head synthesis. We will release the source code upon publication.

Unmanned aerial vehicles (UAVs)-assisted edge computing has emerged as an effective solution for providing contingency task offloading services when ground computing infrastructures are insufficient. However, UAVs face challenges in implementing efficient task offloading strategies due to their limited capabilities and the complexity of the privacy offloading problem. To address these challenges, this study constructs a digital twin (DT)-enabled UAV swarm-assisted secure computing model, which considers collaboration of devices, edges, and cloud resources. The model is designed to represent the three-tier computing environment as a DT virtual framework, allowing for the monitoring of network changes and the exploration of potential strategies. Furthermore, a joint optimization problem that considers time delay and energy consumption within encryption and decryption costs is formulated. To solve this problem, an entropy-enhanced proximal policy optimization-based multi-UAV assisted security-aware task offloading (EP-MUSTO) algorithm is proposed. In EP-MUSTO, the exploration capability is enhanced by utilizing an actor network with policy entropy, and the action cognition is improved through the parameterization of the hybrid action space. Experimental results demonstrate that compared with other advanced algorithms, EP-MUSTO achieves a reduction in security system costs and magnitude of convergence oscillations by at least 9.43% and 54.62%, respectively.

The proliferation of vulnerable code poses a significant threat to software system security and user privacy. Given the inefficiency inherent in manual vulnerability analysis, there has been a pronounced surge of interest in automating vulnerability management using machine learning techniques. However, the scarcity of publicly accessible and large-scale datasets in the vulnerability domain impedes the advancement of automated methodologies. The advent of federated learning has introduced the potential utilization of private data for learning, while ensuring privacy and security within this paradigm presents a novel challenge. To solve this problem, we introduce a new approach called vulnerability solution with abstract syntax tree (AST), SOEHash, and clustering (V-ASC). We first obtain the AST of the vulnerability code to obtain the underlying pattern of the vulnerability. To protect data privacy as well as to extract vector features of the, we use the SOEHash algorithm to process the AST. Finally, to speed up the process of similarity comparison between vectors, we use an unsupervised clustering algorithm to transform the set of vectors into individual vulnerability clusters. Experiments on a recent vulnerability code dataset validate the effectiveness and efficiency of V-ASC.

The vision-based semantic scene completion task aims to predict dense geometric and semantic 3D scene representations from 2D images. However, the presence of dynamic objects in the scene seriously affects the accuracy of the model inferring 3D structures from 2D images. Existing methods simply stack multiple frames of image input to increase dense scene semantic information, but ignore the fact that dynamic objects and non-texture areas violate multi-view consistency and matching reliability. To address these issues, we propose a novel method, CDScene: Vision-based 3D Semantic Scene Completion via Capturing Dynamic Representations. First, we leverage a large multi-modal model to extract 2D explicit semantics and align them into 3D space. Second, we exploit the characteristics of monocular and stereo depth to decouple scene information into dynamic and static features. The dynamic features contain structural relationships around dynamic objects, and the static features contain dense contextual spatial information. Finally, we design a dynamic-static adaptive fusion module to effectively extract and aggregate complementary features, achieving robust and accurate semantic scene completion in autonomous driving scenarios. Extensive experimental results on the SemanticKITTI, SSCBench-KITTI360, and SemanticKITTI-C datasets demonstrate the superiority and robustness of CDScene over existing state-of-the-art methods.

Multi-view bipartite graph clustering (MVBGC) is an active pipeline in unsupervised learning to tackle the limited scalability issue of traditional graph clustering. Despite improved performance, numerous variants still fall under conventional modeling that plugs additional modules, which however induces increasingly intricate models and fails to reveal the inherent variable relationship. We make the first attempt to introduce probabilistic graphical models for modeling the multi-view bipartite graph clustering task, reformulating it as a maximum likelihood estimation (MLE) problem. Such a setting uncovers the underlying probabilistic correlations among the commonality, view-specific variables, and noisy components. By pruning redundancy and disturbance collectively referred to as noise, we prove that minimizing the total noise is an approximation of the lower bound of MLE for multi-view data observations. We further generalize the MLE setting with clustering-suited constraints, deriving a Generalized Probabilistic Graphical Modeling framework (GProM), achieving an interpretable, concise, and flexible MVBGC framework. Extensive experiments verify the effectiveness of our framework. Furthermore, statistical significance analysis reveals the effectiveness of different distribution assumptions, providing valuable insights for model design.

Ring all-reduce is currently the most commonly used collective communication technique in the fields of data parallel and distributed computing. It consists of three phases: communication establishment, data transmission, and data processing at each step. However, this method may suffer from increased communication latency as the number of computation nodes increases, excessive communication steps and data processing procedures can lead to insufficient bandwidth utilization. To address this issue, this article proposes an Interleaved Bidirectional Ring (IBing) all-reduce method, which uses specially crafted communication operations to improve communication efficiency by reducing the effects of both communication establishment and data processing time. IBing reduces the number of communication steps by half compared to the Ring all-reduce. The results of extensive experiments indicate that the proposed IBing design can reduce total communication consumption by an average of 8.49% and up to 49.73%.

Oblivious RAM (ORAM) is an efficient cryptographic primitive that prevents leakage of memory access patterns. It has been referenced by modern secure processors and plays an important role in memory security protection. Although the most advanced ORAM has made great progress in performance optimization, the access overhead (i.e., data blocks) and on-chip (i.e., PosMap) storage overhead is still too high, which will lead to problems such as low system performance. To overcome the above challenges, in this paper, we propose a DC-ORAM system, which reduces the data access overhead and on-chip PosMap storage overhead by using dynamic compression technology. Specifically, we use byte stream redundancy compression technology to compress data blocks on the ORAM tree. And in PosMap, a high-bit multiplexing strategy is used to achieve data compression for binary high-bit repeated data of leaf labels (or path labels). By introducing the above compression technology, in this work, compared with conventional Path ORAM, the compression rate of the ORAM tree is$52.9\%$, and the compression rate of PosMap is$40.0\%$. In terms of performance, compared to conventional Path ORAM, our proposed DC-ORAM system reduces the average latency by$33.6\%$. In addition, we apply the compression technology proposed in this work to the Ring ORAM system. By comparison, it is found that with the same compression ratio as Path ORAM, our design can still reduce latency by an average of$21.5\%$.

Graph Neural Networks (GNNs) excel in processing graph-structured data, making them attractive and promising for tasks such as recommender systems and traffic forecasting. However, GNNs’ irregular computational patterns limit their ability to achieve low latency and high energy efficiency, particularly in edge computing environments. Current GNN accelerators predominantly focus on value sparsity, underutilizing the potential performance gains from bit-level sparsity. However, applying existing bit-serial accelerators to GNNs presents several challenges. These challenges arise from GNNs’ more complex data flow compared to conventional neural networks, as well as difficulties in data localization and load balancing with irregular graph data.To address these challenges, we propose BEAST-GNN, a bit-serial GNN accelerator that fully exploits bit-level sparsity. BEAST-GNN introduces streamlined sparse-dense bit matrix multiplication for optimized data flow, a column-overlapped graph partitioning method to enhance data locality by reducing memory access inefficiencies, and a sparse bit-counting strategy to ensure balanced workload distribution across processing elements (PEs). Compared to state-of-the-art accelerators, including HyGCN, GCNAX, Laconic, GROW, I-GCN, SGCN, and MEGA, BEAST-GNN achieves speedups of 21.7 ×, 6.4×, 10.5×, 3.7×, 4.0×, 3.3×, and 1.4× respectively, while also reducing DRAM access by 36.3×, 7.9×, 6.6×, 3.9×, 5.38×, 3.37×, and 1.44×. Additionally, BEAST-GNN consumes only 4.8%, 12.4%, 19.6%, 27.7%, 17.0%, 26.5%, and 82.8% of the energy required by these architectures.

Sparsification technology is crucial for deploying convolutional neural networks in resource-constrained environments. However, the efficiency of sparse models is hampered by irregular memory access patterns in sparse matrix multiplication kernels. Hardware-level support for 2:4 granularity in sparse tensor cores presents an opportunity for designing efficient sparse matrix multiplication kernels. Existing approaches often involve adjusting sparse structures or secondary sparsification, introducing additional computational errors. To tackle this challenge, we introduce a flexible 2:4 structured adaptive sparse matrix multiplication (FS-AMM) method, a hardware-friendly sparse matrix multiplication kernel that leverages model context to accelerate convolutional neural networks. First, we propose a model context-aware matrix pre-processing method that employs heuristic algorithms to estimate a loss of accuracy due to weight sparsity at each layer. Second, we design a hardware-friendly sparse storage format that combines 2:4 sparse and dense storage formats, enabling more versatile sparsity ratio selection. Third, we implement efficient matrix multiplication kernels to optimize GPU utilization. Finally, experimental results on A100 GPUs show that our method effectively utilizes the sparse tensor kernel and obtains an average 3.09 times speedup ratio compared to other sparse methods while maintaining a high accuracy.

Control flow is the execution order of individual statements, instructions, or function calls within an imperative program. Malicious operation of control flow (e.g., tampering with normal function addresses) leads to severe consequences such as data leakage and system crash. Control Flow Integrity (CFI) is a defense restricting the execution order of program within Control Flow Graph (CFG). IndexHooks is an existing CFI solution designed against forward function calls tampering (including direct and indirect jump). This solution constructs a read-only linear function table that stores function addresses during compilation. Then, IndexHooks checks the table to make program jump to the correct target address during runtime. However, IndexHooks faces limitations in backtracking CFG construction, which can lead to excessive memory usage; the linear structure of the function table is vulnerable to brute force tampering. Addressing the limitations of IndexHooks, this study develops an obfuscated CFI solution called LeafHooks. LeafHooks is implemented during compilation by the LLVM compiler, which performs static analysis and instrumentation on the LLVM Intermediate Representation (IR) code of a program. We make the following three innovations: 1) we propose a speculation-free identification method for indirect function calls by linear traversing and analyzing codes to obtain legal function information (function address); 2) we save this information into a function table in the form of a Balanced Binary Tree (also known as AVL), enhancing the fuzzification of function addresses to defend against brute force; 3) we design a method to simulate control tamper attacks on ARM64 architecture to verify the ability of LeafHooks to protection. LeafHooks shows less overhead than state-of-the-art solutions and reduces 2.9% and 0.55% overhead on average using UnixBench and Phoronix, respectively.

Memory compression can reduce the size of the inactive data in the random access memory (RAM), thereby freeing up unused space and allowing more programs to run; however, current mainstream memory compression frameworks (e.g., Zram and Zswap) and algorithms (e.g., Zstd and Lz4) do not effectively solve the problem of increased CPU utilization, causing they cannot be directly applied to the resource-sensitive embedded operating system, that is, sensitive to both CPU utilization and memory usage. In this study, we develop a Zram instance pool framework called ZramPool for adaptive memory compression. The framework consists of the swap space with multiple Zram instances and the adaptive Zram compression module. Through introducing linear regression analysis, the number of Zram instances can be adaptively adjusted based on the size of the compressed data, allowing Zram instances to work in parallel to match the workload. In ZramPool, we achieve two different requirements of reducing CPU utilization while keeping compression speed and increasing compression speed while keeping CPU utilization. ZramPool is deployed in the embedded Linux OS with a 8GB memory size running on the ARMv8 architecture. For the first requirement, ZramPool can reduce CPU utilization by an average of 11.42% while the compression speed only decreases by an average of 2.4%. For the second requirement, ZramPool can increase compression speed by an average of 11.71% while the CPU utilization only increases by an average of 1.9%.

Point cloud neural networks are gaining increasing attention in emerging 3-D computer vision applications, such as autonomous driving, robotics, and virtual reality. Many customized accelerators for 3-D point clouds have been developed to pursue superior time and energy efficiencies. In this work, we reveal that spatially adjacent points in a 3-D point cloud show similar feature values and relationships, implying substantial redundant computations and memory accesses, while which have been previously ignored. To reduce such redundancies, we propose SimDiff, an algorithm-accelerator co-design framework that boosts 3-D point cloud processing by cleverly leveraging spatial similarity toward excellent speedup and energy efficiency. On the algorithm side, we design a novel similarity-aware differential point cloud neural network (dubbed SD-PCNet). Differing from the standard flow of mainstream point cloud networks, it abstracts a brand-new execution flow for point cloud processing by utilizing spatial similarity among points and dynamic differential execution. On the accelerator side, we propose SD-PCAcc, a supporting accelerator to convert algorithm-level redundancy reductions into performance enhancements. On the deployment side, we propose efficient strategies for network-to-accelerator mapping and scheduling, high-bandwidth memory (HBM) channel allocation, and core component reconfiguration, facilitating the proposed methodologies into practical implementation. Extensive evaluation results show that, with preserved accuracy, our SimDiff gains an average of$3.2\times $speedup and$3.1\times $energy efficiency compared to the state-of-the-art competitors.

System virtualization techniques are gradually applied to embedded systems with the performance improvement of embedded devices; however, there is a large time offset (i.e., deviation) of hundreds of ms between virtual machines (VMs) due to the drift of the virtualization abstraction layer and the jitter phenomenon. It is urgent to significantly reduce this offset by time synchronization techniques. Unfortunately, the current mainstream time synchronization software (e.g., Chrony and LinuxPTP) cannot be used in the embedded virtualized system connected by the fieldbus, such as controller area network (CAN); meanwhile, drift correction-based time synchronization techniques dedicated to CAN bus cannot be utilized for virtualized systems. In this study, we design a general vmPTP, which is a precise time protocol (PTP) for inter-VM communication in embedded virtualized systems based on Linux kernel-based VM (KVM). This is the first time synchronization work focused on embedded virtualized systems. We implement two vmPTP versions based on VirtIO and inter-VM shared memory (IVSHMEM) according to the four message (i.e., packet) exchanges of PTP. We propose an asymmetry compensation strategy to solve the problem of asymmetric rate of I/O operations in the message exchange process, which optimizes the offset computation. We further produce a convergence stopping approach which can continuously converge the offset until the offset is within the given bound and stop. We conduct experimental evaluations on both X86-64 and ARM64 architectures. For two versions of vmPTP, the experimental results show that: 1) the offsets are always within$1~{\mu }$s on two architectures; 2) the offsets are less than that of LinuxPTP and Chrony; and 3) the vmPTP’s CPU and memory utilization on the X86-64 architecture are merely 0.3% and 0.1%, respectively, whereas those on the ARM64 architecture are 3.8% and 0.1%, respectively. We experiment with vmPTP in real-time application and the result demonstrates that vmPTP is feasible and reliable in the real platform.

Stream processing is integral to edge computing due to its low-latency attributes. Nevertheless, variability in user group sizes and disparate computing capabilities of edge devices necessitate frequent operator migrations within the stream. Moreover, intricate dependencies among stream operators often obscure the detection of potential bottleneck operators until an identified bottleneck is migrated in the stream. To address this, we propose a Cost-Aware Operator Migration (CAOM) scheme. The CAOM scheme incorporates a bottleneck operator detection mechanism that directly identifies all bottleneck operators based on task running metrics. This approach avoids multiple consecutive operator migrations in complex tasks, reducing the number of task interruptions caused by operator migration. Moreover, CAOM takes into account the temporal variance in operator migration costs. By factoring in the fluctuating data generation rate from data sources at different time intervals, CAOM selects the optimal start time for operator migration to minimize the amount of accumulated data during task interruptions. Finally, we implemented CAOM on Apache Flink and evaluated its performance using the WordCount and Nexmark applications. Our experiments show that CAOM effectively reduces the number of necessary operator migrations in tasks with complex topologies and decreases the latency overhead associated with operator migration compared to state-of-the-art schemes.

Understanding how the brain responds to external stimuli and decoding this process has been a significant challenge in neuroscience. While previous studies typically concentrated on brain-to-image and brain-to-language reconstruction, our work strives to reconstruct gestures associated with speech stimuli perceived by brain. Unfortunately, the lack of paired {brain, speech, gesture} data hinders the deployment of deep learning models for this purpose. In this paper, we introduce a novel approach, fMRI2GES, that allows training of fMRI-to-gesture reconstruction networks on unpaired data using Dual Brain Decoding Alignment. This method relies on two key components: (i) observed texts that elicit brain responses, and (ii) textual descriptions associated with the gestures. Then, instead of training models in a completely supervised manner to find a mapping relationship among the three modalities, we harness an fMRIto- text model, a text-to-gesture model with paired data and an fMRI-to-gesture model with unpaired data, establishing dual fMRI-to-gesture reconstruction patterns. Afterward, we explicitly align two outputs and train our model in a self-supervision way. We show that our proposed method can reconstruct expressive gestures directly from fMRI recordings. We also investigate fMRI signals from different ROIs in the cortex and how they affect generation results. Overall, we provide new insights into decoding co-speech gestures, thereby advancing our understanding of neuroscience and cognitive science.

Detailed system operations are recorded in logs. To ensure system reliability, developers can detect system anomalies through log anomaly detection. Log parsing, which converts semi-structured log messages into structured data, is a crucial step in log anomaly detection and advanced program analysis and verification. Despite the availability of various log parsing tools, they generally suffer from low parsing accuracy and slow efficiency due to the ignorance of variable characteristics and the use of costly pairwise comparison methods. In this paper, we propose a TCMS framework to parse logs, consisting of two main technologies. First, by studying 16 public log datasets, we find that most log variable tokens are structured variable tokens. Based on this discovery, we propose a token conversion algorithm to improve parsing accuracy. This algorithm converts the changed parts in structured variable tokens into wildcards (‘$< *> $’), preventing these tokens from being directly identified as constant tokens. Second, to improve efficiency, we propose the LogMLCS algorithm, which intelligently constructs a graph to facilitate the extraction of common parts from multiple log messages at once, instead of using pairwise comparisons. Comprehensive experiments conducted on 16 log datasets reveal that our TCMS outperforms seven other parsing methods, achieving the highest parsing accuracy at the fastest speed. Furthermore, experimental results from running a log anomaly detection algorithm in conjunction with different log parsing methods demonstrate that TCMS significantly boosts detection accuracy. For instance, on the OpenStack dataset, our TCMS-facilitated log anomaly detection algorithm achieves a perfect F1-score, precision, and recall of 100% each, surpassing the best peer method by 32.2, 0.8, and 19.5 percentage points, respectively.

Constraints may scatter the Pareto optimal solutions of a constrained multiobjective optimization problem (CMOP) into multiple feasible regions. To avoid getting trapped in local optimal feasible regions or a part of the global optimal feasible regions, a constrained multiobjective evolutionary algorithm (CMOEA) should consider both the escape force and the expansion force carefully during the search process. However, most CMOEAs fail to provide these two forces effectively. As a remedy for this limitation, this article proposes a method called TPEA. TPEA maintains three populations, termed Pop1, Pop2, and Pop3. Pop1 is a regular population, updated with a constrained NSGA-II variant. Pop2 and Pop3 are two auxiliary populations, containing the innermost and outermost nondominated infeasible solutions, respectively. The analysis reveals that these two types of nondominated infeasible solutions can contribute to the generation of escape and expansion forces, respectively. Due to these two forces, TPEA is likely to identify more global optimal feasible regions, which is crucial for constrained multiobjective optimization. Also, a mating selection strategy is developed in TPEA to coordinate the interaction among these three populations. Extensive experiments on 58 benchmark CMOPs and 35 real-world ones demonstrate that TPEA is significantly superior or comparable to six state-of-the-art CMOEAs on most test instances.

Both truth discovery and pattern analysis are effective methods for extracting valuable insights from data streams in mobile crowdsensing. However, existing privacy-preserving schemes either suffer from low data utility or provide high utility at the cost of weak privacy protection. To address this challenge, we introduce a robust privacy-preserving scheme that facilitates high-utility truth discovery and pattern analysis over mobile crowdsensing data streams. Concretely, we leverage the Square Wave mechanism, a randomized reporting technique, to perturb the data to prevent privacy breaches. To reduce the utility loss caused by perturbation, we design a budget allocation algorithm. This algorithm ensures that adjacent timestamps with approximate data share a perturbed value derived from their accumulated budgets. Furthermore, to facilitate robust pattern analysis, we propose a data splitting method that divides the perturbed data into two parts: one part records patterns randomly, while the other part recovers the perturbed values. Theoretical analysis confirms that our scheme satisfies ω-event ϵ-differential privacy level. Extensive experiments conducted on four real-world datasets demonstrate that our scheme outperforms existing schemes, delivering more accurate results for both truth discovery and pattern analysis under the same privacy constraints.

In online social networks (OSNs), people usually join groups for communication. Information diffusion often occurs with some cost, either between individuals or within/among groups; and different opinions may compete with each other. The groups can make decisions based on the majority of the group members. This type of group consensus is common in group activities. However, existing research on maximization of competitive influence often neglects the effects of group consensus. To this end, we introduce the process of group consensus reaching in influence maximization and propose a novel Group consensus-based Competitive Linear Threshold (GCLT) propagation model; then we study the Budgeted Competitive Influence Maximization (BCIM) problem under the GCLT model. We reveal that the problem is NP-hard, and the objective function is proven to be neither submodular nor supermodular. To this end, we construct an equivalent Group consensus-based Competitive Live-Edge (GCLE) model of GCLT by sampling method. Based on GCLE, develop two submodular functions of the upper and lower bounds. Then, we propose the SBG algorithm by applying the Sandwich Approximation framework for the BCIM problem under the GCLT model. In SBG , we provide an approximate solution to the lower bound and the upper bound by the proposed OPIM-B algorithm. Then, we select the seed set of solutions that achieves the best influence spread in Monte Carlo simulations. We also propose two strategies to optimize SBG . The experiments on six real social network datasets verify the effectiveness and scalability of our method and validate the impact of group consensus on the competitive influence dissemination process, as well as the importance of considering the process of reaching group consensus.

Truth discovery endeavors to extract valuable information from multi-source data through weighted aggregation. Some studies have integrated differential privacy techniques into traditional truth discovery algorithms to protect data privacy. However, due to the neglect of outliers and limitations in budget allocation, these schemes still need improvement in the accuracy of discovery results. To solve these challenges, we propose a privacy-preserving scheme called PriPTD to achieve secure and accurate truth discovery services over crowdsourced data streams. Instead of assuming that worker weights are always stable between two neighboring timestamps, we delve deeper to consider outliers where worker weights change rapidly. Accordingly, we develop an outlier-aware weight estimation method with a time series model to capture and handle these outliers. Furthermore, to ensure data utility under a limited budget, we devise a weight-aware budget allocation algorithm. Its core idea is that timestamps with higher importance consume a larger proportion of the remaining budget. Additionally, we design a noise-aware error adjustment approach to mitigate the adverse effects of introduced noise on accuracy. Theoretical analysis and extensive experiments validate our scheme. Final comparative experiments against existing works confirm that our scheme achieves more accurate truth discovery while preserving privacy.

Bipartite graph clustering (BGC) has emerged as a fast-growing research in the clustering community. Despite BGC has achieved promising scalability, most variants still suffer from the following concerns: a) Susceptibility to noisy features. They construct bipartite graphs in the raw feature space, inducing poor robustness to noisy features. b) Inflexible anchor selection strategies. They usually select anchors through heuristic sampling or constrained learning methods, degrading flexibility. c) Partial structure mining. Existing methods are mainly built upon Linear Reconstruction Paradigm (LRP) from subspace clustering or Locally Linear Paradigm (LLP) from manifold learning, which partially exploit linear or locally linear structures, lacking a unified perspective to integrate global complementary structures. To this end, we propose a novel model, termedJoint Robust Embedding and Structural FusionBipartiteGraphClustering (JetBGC), which focuses on three aspects, namely robustness, flexibility, and complementarity. Concretely, we first introduce a robust embedding learning module to extract latent representation that can reduce the impact of noisy features. Then, we optimize anchors via a constraint-free strategy that can flexibly capture data distribution. Furthermore, we revisit the consistency and specificity of LRP and LLP, and design a new unified structural fusion strategy to integrate both linear and locally linear structures from a global perspective. Therefore, JetBGC unifies robust representation learning, flexible anchor optimization, and structural bipartite graph fusion in a framework. Extensive experiments on synthetic and real-world datasets validate our effectiveness against existing baselines.

Sparse data gathering has become a promising solution for reducing measurement costs by leveraging the inherent sparsity of data. However, most existing approaches rely on low-dimensional models such as compressive sensing or matrix completion, which are limited in capturing complex high-dimensional structures. To overcome these limitations, we proposeTensorMon, a novel tensor-based sparse data gathering framework that introduces a cuboid sampling strategy to more effectively exploit multidimensional correlations. Unlike traditional entry-based or tube-based sampling, TensorMon introduces the innovative concept ofcuboid sampling. We further develop a lightweight sampling scheduling algorithm and a non-iterative inference algorithm to ensure efficient measurement planning and accurate reconstruction of unmeasured data. Theoretical analysis establishes a new performance bound for our sampling strategy, which is significantly lower than those in existing literature. To validate our theoretical findings, we conduct extensive experiments on four real-world datasets: two network monitoring datasets, a city-scale crowd flow dataset, and a road traffic speed dataset. Experimental results demonstrate that TensorMon achieves substantial reductions in measurement cost, delivers high inference accuracy, and ensures rapid data recovery, highlighting its effectiveness and practicality across diverse application scenarios.

This paper addresses the pressing need for effective k-tips decomposition in dynamic bipartite graphs, a crucial aspect of real-time applications that analyze and mine binary relationship patterns. Recognizing the dynamic nature of these graphs, our study is the first to provide a solution for k-tips decomposition in such evolving environments. We introduce a pioneering projection-based algorithm, coupled with advanced incremental maintenance strategies for edge modifications, tailored specifically for dynamic graphs. This novel approach not only fills a significant gap in the analysis of dynamic bipartite graphs but also substantially enhances the accuracy and timeliness of data-driven decisions in critical areas like public health. Our contributions set a new benchmark in the field, paving the way for more nuanced and responsive analyses in various domains reliant on dynamic data interpretation.

Machine unlearning involves retracting data records and reducing their influence on trained models, aiding user privacy protection, at a significant computational cost potentially. Weight perturbation-based unlearning is common but typically modifies parameters globally. We propose fine-grained Top-K and Random-k parameters perturbed inexact machine unlearning that address the privacy needs while keeping the computational costs tractable. However, commonly used training data are independent and identically distributed, for inexact machine unlearning, current metrics are inadequate in quantifying unlearning degree that occurs after unlearning. To address this quantification issue, we introduce SPD-GAN, which subtly perturbs data distribution targeted for unlearning. Then, we evaluate unlearning degree by measuring the performance difference of the models on the perturbed unlearning data before and after unlearning. Furthermore, to demonstrate efficacy, we tackle the challenge of evaluating machine unlearning by assessing model generalization across unlearning and remaining data. To better assess the unlearning effect and model generalization, we propose novel metrics, namely, the forgetting rate and memory retention rate. By implementing these innovative techniques and metrics, we achieve computationally efficacious privacy protection in machine learning applications without significant sacrifice of model performance. A by-product of our work is a novel method for evaluating and quantifying unlearning degree.

Graph data analysis has been used in various real-world applications to improve services or scientific research, which, however, may expose sensitive personal information. Differential privacy (DP) has become the gold standard for publishing graph data while still protecting personal privacy. However, most existing studies over differentially private graph data publication mainly focus on static unweighted graphs. As interactions between entities in real systems are often dynamically changing and associated with weights, it is desirable to consider the more general scenario of continuous weighted graph publication under DP in the temporal dimension. Therefore, we investigate the problem of publishing weighted graphs satisfying DP under continuous monitoring. Specifically, we consider a server that continuously monitors user data and publishes a sequence of weighted graph snapshots. We propose SwgDP, a novel framework that leverages historical graph data to guide current snapshot generation. SwgDP consists of four key components: node adaptive sampling, dynamic weight optimization, prediction-based community detection and weighted graph generation. We demonstrate that SwgDP satisfies DP, and comprehensive experiments on four real-world datasets and four commonly used graph metrics show that SwgDP can effectively synthesize weighted graph at any time step.

The detection of cardiac phase in ultrasound videos, identifying end-systolic (ES) and end-diastolic (ED) frames, is a critical step in assessing cardiac function, monitoring structural changes, and diagnosing congenital heart disease. Current popular methods use recurrent neural networks to track dependencies over long sequences for cardiac phase detection, but often overlook the short-term motion of cardiac valves that sonographers rely on. In this paper, we propose a novel optical flow-enhanced Mamba U-net framework, designed to utilize both short-term motion and long-term dependencies to detect the cardiac phase in ultrasound videos. We utilize optical flow to capture the short-term motion of cardiac muscles and valves between adjacent frames, enhancing the input video. The Mamba layer is employed to track long-term dependencies across cardiac cycles. We then develop regression branches using the U-Net architecture, which integrates short-term and long-term information while extracting multi-scale features. Using this method, we can generate regression scores for each frame and identify keyframes (i.e., ES and ED frames). Additionally, we design a keyframe weighted loss function to guide the network to focus more on keyframes rather than intermediate period frames. Our method demonstrates superior performance compared to advanced baseline methods, achieving frame mismatches of 1.465 frames for ES and 0.842 frames for ED in the Fetal Echocardiogram dataset, where heart rates are higher and phase changes occur rapidly, and 2.444 frames and 2.072 frames in the publicly available adult Echonet-Dynamic dataset. Its accuracy and robustness in both fetal and adult datasets highlight its potential for clinical application.

This brief is concerned with the prediction problem of product popularity under a social network (SN) with positive-negative diffusion (PND). First, a PND model is proposed to enable the simulation of product diffusion, and three user states are defined. Second, an optimal and precise feature vector of every user is extracted through a multi-agent-system-based attention mechanism (MASAM) that is devised. The weight matrix shared in the mechanism of all agents is learned using a distributed learning algorithm provided in MASAM. Third, an MAS model for product diffusion on SN is established based on the feature representations from MASAM. Rules for agent interaction during PND diffusion are suggested, which accelerate the simulation of information spread in SN. Finally, comprehensive experiments are conducted to verify the effectiveness and efficiency of the proposed models and algorithms in prediction and to compare their performance with baseline methods. Furthermore, a case study is provided to illustrate the applicability and extendibility of the developed algorithm.

This article addresses the issue of multiple influences maximization under dynamic link strength (MIMDLS) in multi-agent systems (MASs). Initially, a novel model for dynamic link strength within MASs is suggested to facilitate the simulation of multiple influences diffusion. Subsequently, the MIMDLS problem is formulated with both competitive and cooperative scenarios being examined. In response, two diffusion models, specifically the competitive multiple influences independent cascade (Cp-MIIC) model and the cooperative multiple influences linear threshold (Cr-MILT) model, are designed for MASs. Furthermore, a distributed deep reinforcement learning (DRL) framework is established based on MASs by incorporating asynchronous training and updating processes for seed selection in the context of multiple influences. Moreover, the developed distributed DRL algorithm encompasses the estimation of Q value as well as the management of constraints within Cp-MIIC and Cr-MILT models. Finally, comprehensive experiments are conducted to: 1) validate the effectiveness and efficiency of the proposed models and algorithms in terms of multiple influence diffusion and 2) benchmark their performance against state-of-the-art methods.

Video instance segmentation (VIS) is a challenging task, requiring handling object classification, segmentation, and tracking in videos. Existing Transformer-based VIS approaches have shown remarkable success, combining encoded features and instance queries as decoder inputs. However, their decoder inputs are low-resolution due to computational cost, resulting in a loss of fine-grained information, sensitivity to background interference, and poor handling of small objects. Moreover, the queries are randomly initialized without location information, hindering convergence efficiency and accurate object instance localization. To address these issues, we propose a novel VIS approach, STFormer, with a spatial-temporal feature aggregation (STFA) module and spatial-temporal-aware Transformer (STT). Specifically, STFA obtains robust high-resolution masked features efficiently for the decoder, while STT's location-guided instance query (LGIQ) improves initial instance queries. STFormer preserves more fine-grained information, improves convergence efficiency, and localizes object instance features accurately. Extensive experiments on YouTube-VIS 2019, YouTube-VIS 2021, and OVIS datasets show that STFormer outperforms mainstream VIS methods.

Deep neural networks (DNNs) have made great breakthroughs and seen applications in many domains. However, the incomparable accuracy of DNNs is achieved with the cost of considerable memory consumption and high computational complexity, which restricts their deployment on conventional desktops and portable devices. To address this issue, low-rank factorization, which decomposes the neural network parameters into smaller sized matrices or tensors, has emerged as a promising technique for network compression. In this article, we propose leveraging the emerging tensor ring (TR) factorization to compress the neural network. We investigate the impact of both parameter tensor reshaping and TR decomposition (TRD) on the total number of compressed parameters. To achieve the maximal parameter compression, we propose an algorithm based on prime factorization that simultaneously identifies the optimal tensor reshaping and TRD. In addition, we discover that different execution orders of the core tensors result in varying computational complexities. To identify the optimal execution order, we construct a novel tree structure. Based on this structure, we propose a top-to-bottom splitting algorithm to schedule the execution of core tensors, thereby minimizing computational complexity. We have performed extensive experiments using three kinds of neural networks with three different datasets. The experimental results demonstrate that, compared with the three state-of-the-art algorithms for low-rank factorization, our algorithm can achieve better performance with much lower memory consumption and lower computational complexity.

The serendipity recommendation tries to burst the filter bubble while still meeting user interests. However, serendipity itself has not been well understood in the recommendation system. Thus, factor investigation in recommendation serendipity has attracted much attention, for which two challenges hinder follow-up research: (1) Ambiguity of factors . Different works exploit different factors, and the meanings of factors are inconsistent in various works. (2) Lack of complete impact validation . The importance of these factors in different domains is not yet fully understood. The common approach of user surveys costs much, but the results are usually less objective and limited in quantity. To this end, we strive to comprehensively identify and clarify serendipity factors and explore objective data-driven approaches to validate factor impacts in large-scale cross-domain scenarios. We first conduct a comprehensive literature review to identify all possible factors, from which we find that some factors are being used indistinguishably. To address this issue, we propose two principles of meaning coverage and factor independence to clarify and disentangle serendipity factors. Next, we propose a general experimental framework to explore the impacts of factors. Then, we implement one such framework and run experiments on nine representative datasets to study factor importance on serendipity. We also propose a quantitative method to measure the degree of disentanglement of factors and to test the effects of factor combinations. We gain several useful findings: (1) relevance , diversity , and random are critical factors affecting serendipity; (2) domain features affect factor importance and can guide serendipity recommendation; (3) the disentanglement quantification method benefits the understanding of serendipity and the combination of factors. To our knowledge, this is the first work to comprehensively investigate serendipity factors and experimentally compare their impacts in an objective data-driven approach.

Network distance measurement is crucial for evaluating network performance, attracting significant research attention. However, conducting measurements for the entire network is exceedingly expensive and time-consuming, making the reduction of network distance measurement costs a top priority. The tensor completion method efficiently reduces measurement costs by utilizing a small amount of measured data to estimate the entire network’s distance data. Unfortunately, current tensor completion methods still suffer from issues such as complex sample selection, high measurement overhead, slow recovery, and low inference accuracy. To address the aforementioned challenges, we present an online network-wide distance measurement scheme. In this approach, continuous distance data are structured into sliding-window-based tensors. Our method incorporates a lightweight sample selection algorithm with a lowest sampling bound and a rapid, accurate unmeasured data inference algorithm. We have conducted extensive experiments using four real network distance datasets and two citywide crowd flow datasets. The empirical evaluations demonstrate the effectiveness of our approach, particularly in reducing measurement costs and enhancing data recovery accuracy.

5G mobile communication techniques are widely applied in ubiquitous communication scenarios (e.g., smart navigation and smart transportation), which require time synchronization among mobile devices. However, the built-in time synchronization software of Android phones presents large time offsets with hundreds of milliseconds (ms), and the mainstream time synchronization techniques have specific limitations: 1) the Network Time Protocol (NTP) has too large offset to meet the real-time information interaction among mobile devices; 2) the Linux Precision Time Protocol (LinuxPTP) exists hardware dependence and cannot be implemented on Android and 5G networks; and 3) the Global Navigation Satellite System (GNSS) requires installing a hardware receiver on each mobile device. In this study, we develop a Mobile Precision Time Protocol (MobiPTP), which is hardware-independent and compatible with various network types, including Wide-Area Network (WAN), Local-Area Network (LAN), wired and wireless networks. The main challenges include signal strength instability and uplink-downlink asymmetry. We propose a dynamic time synchronization algorithm and an asymmetry compensation strategy to overcome these challenges. Regardless of high-speed mobile or crowded conditions in 5G networks, MobiPTP demonstrates an average offset of 9 ms, outperforming the NTP-based open-source software Chrony (about 30 ms). MobiPTP has been successfully deployed in multiple real-world ubiquitous communication scenarios and always demonstrates much lower offsets than Chrony.

In the context of distributed deep learning, computational clusters place greater emphasis on static all-reduce communication latency while also needing to support large-scale networking. However, the communication efficacy of current all-reduce algorithms within specialized network topologies requires enhancement. Existing all-reduce communication algorithms inadequately exploit cluster bandwidth, leading to considerable bandwidth idleness. The optimization of communication algorithms becomes imperative to fully utilize the available bandwidth. Addressing this concern, we propose an innovative approach: a topology-aware interleaved all-reduce algorithm for Dragonfly networks (TIAD). Leveraging the inherent characteristics of the Dragonfly network, TIAD employs an interleaved communication mechanism for both intra- and inter-group data collection, significantly augmenting communication efficiency. Moreover, we refine the Dragonfly network with minimal adjustments, aligning it with the theoretical structure of interleaved communication. We also proposed an all-reduce communication method to complement the TIAD algorithm, specifically for scenarios where only a subset of nodes in the Dragonfly network participate in the communication task. Our experiments demonstrate that TIAD exhibits the shortest communication time across diverse node sizes and bandwidth conditions. Notably, our algorithm reduces communication time by up to 23.4% during the collection communication phase in comparison to the PAARD algorithm.

Byte-addressable non-volatile memory (NVM) is a promising alternative technology for main memory, allowing the processor to access persistent data in the main memory directly. Systems with emerging NVM as the main memory still suffer from information leakage, performance, and endurance challenges. Fortunately, Oblivious RAM (ORAM) provides provable secure address obfuscation and encryption and can be used to build a secure memory system. However, it is challenging to support crash consistency with the increasing use of NVM as persistent memory. Current state-of-the-art PS-ORAM systems are designed based on pure NVM memory systems, which suffer from low performance. Different from the previous PS-ORAM system, in this work, we consider a hybrid memory system (both NVM and DRAM) and aim to build a crash-consistent ORAM framework with low overhead. We propose the HM-ORAM framework, which includes a novel ORAM tree-level partition scheme, a lightweight data persistency architecture, and a secure access protocol. It takes full advantage of both hybrid memory technologies to achieve the design requirements to support crash consistency and persistence of ORAM systems with minimal write overhead. Without compromising security, HM-ORAM can outperform an NVM-based ORAM (i.e., NVM-ORAM) system by 1.23× and 1.39× in non-recursive and recursive implementations.

Zero-Knowledge Proofs (ZKP) and Homomorphic Encryption (HE) are crucial for data privacy in applications like cloud, blockchain, and analytics. However, the real-world adoption often faces performance challenges, particularly in the execution of the Number Theoretic Transform (NTT) required for polynomial multiplication involving sizes beyond \(2^{20}\) and large integer widths (e.g., 256 bits). FPGAs offer a promising platform for acceleration, but efficiently implementing large-size NTTs remains difficult due to the limited on-chip resources. The widely adopted four-step NTT method, used to relieve the need for large on-chip memory, introduces performance bottlenecks. Initially, the traditional dataflow NTT architecture may not fully exploit available compute capability, which hinders achieving peak performance. Furthermore, during the matrix transpose phase, the non-sequential access to external High-Bandwidth Memory (HBM) causes inefficiency. To address these challenges, we introduce HiFA, an FPGA-based automatic accelerator framework designed for high-performance and flexible large-size NTT computations. HiFA utilizes a stacked NTT architecture for high parallelism, maximizing HBM throughput. It supports various decomposed polynomial sizes via a novel reordering module. Additionally, a specialized cyclic shuffle module is integrated to optimize data movement during the matrix transpose step, alleviating random memory access delay. HiFA also provides an automatic Design Space Exploration (DSE) framework that identifies optimal four-step decomposition parameters and generates corresponding hardware configurations. Our experiments show that the FPGA implementation of HiFA achieves an average speedup of 2.97× and up to 7.25× improvement in latency over prior state-of-the-art FPGA solutions. Compared to prior GPU-based methods, HiFA achieves an average energy efficiency gain of 2.24×.

Power consumption prediction is one of the keys to optimize the energy consumption of servers. Existing traditional regression-based methods are too simple and poorly generalized, while popular deep learning methods require too much data. Therefore, they are difficult to be widely generalized. In this study, we propose a framework of linear interpolation multi-head sparse temporal pattern attention (Li-MSA) based on few-shot learning for power consumption prediction of servers with small-scale datasets in environments such as cloud data centers or edge computing. First, the interpolation reconstruction module extends and smooths the data. Then, the embedding learning module is used to narrow the scope of the hypothesis space. Finally, the multi-head sparse temporal pattern attention module emphasizes features and predicts power consumption. The results of the experiments show that Li-MSA outperforms the best results among the other methods for two datasets with different time steps in the RMSE metric by 15.34%, 17.35%, 18.18%, 6.28%, 4.05%, 7.73%.

The surge in demand for latency-sensitive services has propelled network distance measurement to the forefront of networking research. Utilizing the low-rank structure of full network data, the tensor completion method can efficiently estimate network distance from partially sampled distance data measured from a small set of node pairs. However, its performance is affected by sampling algorithm limitations, including unreliability and high overhead in dynamic networks. To tackle these challenges, we propose tube-based sampling as an alternative to point-based sampling, utilizing a partition-based algorithm to incorporate randomness for improved reliability. Additionally, we introduce a Tube Length Identification Algorithm to dynamically adjust tube length based on network status, balancing scheduling overhead reduction with estimation accuracy. Experimental results on three real network distance datasets, compared against 13 baseline algorithms, demonstrate the high accuracy and low scheduling overhead of our approach.

Network traffic matrices often suffer from incompleteness and sparsity due to various factors, including network device policies and system limitations. The incompleteness can undermine the reliability and accuracy of network traffic monitoring services, negatively impacting downstream tasks such as network planning and fault diagnosis. Our focus is on network traffic data recovery, intending to infer missing traffic data from partial measurements accurately. Although tensor completion algorithms are quite effective in recovering traffic data, existing models often overlook cross-domain traffic relationships and fail to account for the order and distribution of traffic, leading to reduced recovery accuracy. To overcome these limitations, we propose a new contrastive tensor completion model that utilizes pairwise joint embedding. This model employs innovative techniques, including a cross-domain embedding module to avoid information homogeneity and enhance model expressiveness, a contrastive module to preserve the order and distribution of traffic volumes, and an injective interaction module to map entry embeddings into the numerical space, ensuring convergence and retaining the original numerical distribution. Experiments on three real-world network traffic datasets show that our model significantly reduces the error in missing traffic data recovery compared to other existing models while maintaining traffic order and distribution.

Deep learning has made remarkable progress across many domains, enabled by the capabilities of over-parameterized neural networks with increasing complexity. However, practical applications often necessitate compact and efficient networks because of device constraints. Among recent low-rank decomposition-based neural network compression techniques, Tucker decomposition has emerged as a promising method which effectively compresses the network while preserving the high-order structure and information of the parameters. Despite its promise, designing an efficient Tucker decomposition approach for compressing neural networks while maintaining accuracy is challenging, due to the complexity of setting ranks across multiple layers and the need for extensive fine-tuning. This paper introduces a novel accuracy-aware network compression problem under Tucker decomposition, which considers both network accuracy and compression performance in terms of parameter size. To address this problem, we propose an efficient alternating optimization algorithm that iteratively solves a network training sub-problem and a Tucker decomposition sub-problem to compress the network with performance assurance. The proper Tucker ranks of multiple layers are selected during network training, enabling efficient compression without extensive fine-tuning. We conduct extensive experiments, implementing image classification on five neural networks using four benchmark datasets. The experimental results demonstrate that, without the need for extensive fine-tuning, our proposed method significantly reduces the model size with minimal loss in accuracy, outperforming baseline methods.

Many researchers primarily rely on modeling user interests for sequential recommendation. However, dynamic user behaviors often accompany unstable interaction histories, and modeling user interests alone is insufficient for comprehensive user features. Some studies notice the problem of interest drift, but they are usually limited to modeling at the item-level, unable to perceive the subtle changes at the feature-level. To this end, we propose a Stability-aware Preference Model (SAPM), which consists of three modules. The LSI module for extracting long and short-term interests, the FLC module for extracting feature-level candidate information, and the SAF module for fusing them according to the stability score. In particular, we propose a Multi-head GRU (MHGRU) structure in the LSI module, which is more efficient than the general GRU and has stronger expression ability. Through extensive experiments, our framework shows significant mitigation of the impact of unstable interactions. On the two real data sets, we improve MRR by 5.7% and 14.0% compared with the recent baselines. Moreover, we conduct an in-depth analysis of user interaction stability and obtain several interesting findings that can benefit future studies.

While decentralized training is attractive in multi-agent reinforcement learning (MARL) for its excellent scalability and robustness, its inherent coordination challenges in collaborative tasks result in numerous interactions for agents to learn good policies. To alleviate this problem, action advising methods make experienced agents share their knowledge about what to do, while less experienced agents strictly follow the received advice. However, this method of sharing and utilizing knowledge may hinder the team's exploration of better states, as agents can be unduly influenced by suboptimal or even adverse advice, especially in the early stages of learning. Inspired by the fact that humans can learn not only from the success but also from the failure of others, this paper proposes a novel knowledge sharing framework called Cautiously-Optimistic kNowledge Sharing (CONS). CONS enables each agent to share both positive and negative knowledge and cautiously assimilate knowledge from others, thereby enhancing the efficiency of early-stage exploration and the agents' robustness to adverse advice. Moreover, considering the continuous improvement of policies, agents value negative knowledge more in the early stages of learning and shift their focus to positive knowledge in the later stages. Our framework can be easily integrated into existing Q-learning based methods without introducing additional training costs. We evaluate CONS in several challenging multi-agent tasks and find it excels in environments where optimal behavioral patterns are difficult to discover, surpassing the baselines in terms of convergence rate and final performance.

As artificial intelligence progresses in the field of medical ultrasound image analysis, mitigating catastrophic forgetting during continuous learning processes in disease diagnosis and fetal ultrasound assistance is crucial. Inspired by the significant performance improvements achieved in various downstream visual tasks through advanced visual representations, we introduce a novel approach called Masked Ultrasound Image Modeling (MUIM) to prevent forgetting of new disease diagnosis tasks. In this approach, MUIM initially pre-trains on ultrasound images specific to the current task. By utilizing a masked autoencoder to train on unlabeled ultrasound images, it learns highly abstract representations of task-relevant information. To ensure the model’s adaptability to new tasks, we propose contrastive Historical-Current Learning strategy, which enhances the model’s ability to retain and integrate knowledge from previous tasks while learning new ones. By incorporating knowledge distillation loss and the exponential moving average (EMA) technique for joint inference and continuously updating new knowledge into the historical expert model, our approach enables the model to adaptively learn new tasks while preventing forgetting of old ones. We conducted extensive experiments on three datasets: the Breast Ultrasound Image dataset (BUSI), the Algeria Thyroid Ultrasound Image dataset (AUITD), and our designed Obstetrics and Gynecology Ultrasound Dataset (USOGD). The results demonstrate that our method significantly reduces the forgetting of old task knowledge while outperforming state-of-the-art methods in classification accuracy.

Diversity is increasingly recognized as a crucial factor in recommendation systems for enhancing user satisfaction. However, existing studies on diversity-enhanced recommendation systems primarily focus on designing recommendation strategies, often overlooking the development of evaluation metrics. Widely used diversity metrics such as CC, ILAD, and ILMD are typically assessed independently of accuracy. This separation leads to a critical limitation: existing diversity measures are unable to distinguish between diversity improvements from effective recommendations and those from in effective recommendations. Our evaluations reveal that the diversity improvements are primarily contributed by ineffective recommendations, which often do not positively contribute to user satisfaction. Furthermore, existing diversity metrics disregard the feature distribution of ground-truth items, potentially skewing the assessment of diversity performance. To address these limitations, we design three new accuracy-aware metrics: DCC, FDCC, and DILAD, and conduct a re-evaluation using these metrics. Surprisingly, our results illustrate that the diversity improvements of existing diversity-enhanced approaches are limited and even negative compared to those of accurate recommendations. This finding underscores the need to explore more sophisticated diversity-enhanced techniques for improving the diversity within effective recommendations.

The anatomical structure detection of fetal cardiac views is crucial for diagnosing fetal congenital heart disease. In practice, there is a large domain gap between different hospitals' data, such as the variable data quality due to differences in acquisition equipment. In addition, accurate annotation information provided by obstetrician experts is always very costly or even unavailable. This study explores the unsupervised domain adaptive fetal cardiac structure detection issue. Existing unsupervised domain adaptive object detection (UDAOD) approaches mainly focus on detecting objects in natural scenes, such as Foggy Cityscapes, where the structural relationships of natural scenes are uncertain. Unlike all previous UDAOD scenarios, we first collected a Fetal Cardiac Structure dataset from two hospital centers, called FCS, and proposed a multi-matching UDA approach (M3-UDA), including Histogram Matching (HM), Sub-structure Matching (SM), and Global-structure Matching (GM), to better transfer the topological knowledge of anatomical structure for UDA detection in medical scenarios. HM mitigates the domain gap between the source and target caused by pixel transformation. SM fuses the different angle information of the sub-structure to obtain the local topological knowledge for bridging the domain gap of the internal sub-structure. GM is designed to align the global topological knowledge of the whole organ from the source and target domain. Extensive experiments on our collected FCS and CardiacUDA, and experimental results show that M3-UDA outperforms existing UDAOD studies significantly. Datasets and source code are available at https://github.com/xmed-lab/M3-UDA.

Camera-based Semantic Scene Completion (SSC) is to infer the full geometry of objects and scenes from only 2D images. The task is particularly challenging for those in-visible areas, due to the inherent occlusions and lighting ambiguity. Existing works ignore the information missing or ambiguous in those shaded and occluded areas, resulting in distorted geometric prediction. To address this issue, we propose a novel method, Bi-SSC, bidirectional geomet-ric semantic fusion for camera-based 3D semantic scene completion. The key insight is to use the neighboring structure of objects in the image and the spatial differences from different perspectives to compensate for the lack of information in occluded areas. Specifically, we introduce a spatial sensory fusion module with multiple association attention to improve semantic correlation in geometric distributions. This module works within single view and across stereo views to achieve global spatial consistency. Experimental results demonstrate that Bi-SSC outperforms state-of-the-art camera-based methods on SemanticKITTI, particularly excelling in those invisible and shaded areas.

With the development of chiplet technology, the architecture of Non-Uniform Memory Access (NUMA) has become increasingly intricate. The placement of memory page significantly influences application performance in NUMA systems. We found that memory access bottlenecks occur between high-level NUMA domains consisting of multiple chiplets. In this paper, we introduce a Traffic-Aware Page Mapping Method (TAPMM) designed for multi-level NUMA systems. TAPMM conceptualizes the multi-level NUMA system as a memory access tree, utilizing hardware performance events to be aware of system traffic and identify the optimal page mapping method for bandwidth efficiency. Our experiments demonstrate that TAPMM achieves a speedup of up to 2.12× on a real commodity machine compared to existing optimization tools.

Compared with conventional SRAM, Spin-Transfer Torque Random Access Memory(STT-RAM) is expected to play a crucial role in future memory technologies with the increasing demands for higher storage density and lower power consumption for modern embedded systems. Moreover, Multi-Level Cell (MLC) STT-RAM outperforms Single-Level Cell (SLC) STT-RAM since it has higher bit density. However, MLC STT-RAM suffers from write performance due to the two-step state transitions (TTs) in memory cells' soft domain. State-of-the-art approaches mitigate this issue by reducing TTs with efficient data coding. Unfortunately, none of the existing works can fully eliminate the TTs. In this work, zeroTT, an optimal (3, 4)-based expansion coding method that eliminates TTs for MLC STT-RAM. The design of ZeroTT considers space overhead and coding complexity, and our experimental results demonstrate that zeroTT can completely avoid TTs, leading to a more efficient MLC STT-RAM memory in terms of access latency, energy consumption, and device lifetime.

Long-term occlusion is one of the most formidable challenges in Multi-Object Tracking (MOT). The motion models of existing SORT-based trackers are unreliable in estimating the motion states of long-term occluded targets. This is mainly because as the occlusion period increases, the increases speed of estimation errors in the motion model increases faster. In practical applications, we believe that the estimation error of the tracker during long-term occlusion is mainly concentrated in the estimation error of the motion model on the velocity of the occluded target. In this work, we have demonstrated that in the long-term occlusion period, appropriately correcting the estimated values of the motion model on the target motion velocity and fully utilizing the temporal and attribute information of the target’s historical trajectory as calculation indicators of correlation are beneficial for improving the robustness of the tracker in long-term occlusion. We refer to our proposed motion correction-based framework as MC-SORT, which mainly consists of a Momentum Compensation Module (MCM) and a Backtracking Re-association (BRA) module. The former can correct the estimated value of the target’s motion state during long-term occlusion, the latter uses the temporal and attribute information of the target’s historical trajectory during long-term occlusion as correlation indicators to measure the degree of correlation between the target and trajectory. Our proposed MC-SORT has the characteristics of simplicity, online, real-time, and plug-and-play, particularly improving the robustness of the tracker in long-term occlusion. The extensive experimental results on the MOT17 and MOT20 datasets demonstrate the robustness and superiority of our framework.

The goal of the Single-Row Facility Layout Problem (SRFLP) is to arrange facilities along a straight line in such a way that the total layout cost is minimized. This cost is calculated as the sum of the products of flow costs and the distances between pairs of facilities. SRFLP has numerous practical applications in real-life scenarios, but it is NP-hard. To solve the Single-Row Facility Layout Problem (SRFLP), we propose a novel solution method SRFL-DP. This approach models SRFLP as an electrostatic equilibrium problem, enabling the derivation of a feasible solution in a very short time. Then, a simulated annealing algorithm is applied to this feasible solution, allowing for the exchange of facilities to further reduce the total layout cost. Our method significantly reduces the computational time required to obtain results. When compared with the state-of-the-art KMPG algorithm on the Ranlarge dataset with 2000 facilities, our method takes only 1/11th of the time and performs even faster on other datasets.

Link prediction in directed graphs offers a solution for uncovering detailed and accurate relationships among distinct entities. Unlike conventional link prediction in undirected graphs, the task becomes more intricate in directed graphs as it involves predicting both associations and orientations. Existing methods simply apply classic graph embedding techniques to learn node representations, followed by mapping representations of corresponding node pairs into probabilities indicating potential links. However, the inadequate capture of orientation information and sole reliance on node representations for prediction hinder the effective differentiation of orientation, thereby impeding the link prediction accuracy. In response, we introduce DGLP, an orientation-aware link prediction method tailored for directed graphs. DGLP utilizes the incidence matrix to learn both node and edge representations, effectively capturing structural and orientation information. By leveraging edge representations, DGLP achieves accurate link prediction in directed graphs without relying solely on implicit node representations. Experiments across six datasets demonstrate the effectiveness of DGLP, achieving a 1.2x improvement in prediction results.

New categories may be introduced over time, or existing categories may need to be reclassified. Class incremental learning (CIL) is employed for the gradual acquisition of knowledge about new categories while preserving information about previously learned ones in such dynamic environments. It might also be necessary to also eliminate the influence of related categories on the model to adapt to reclassification. We thus introduce class-level machine unlearning (MU) within CIL. Typically, MU methods tend to be time-consuming and can potentially harm the model’s performance. A continuous stream of unlearning requests could lead to catastrophic forgetting. To address these issues, we propose a non-destructive eCIL-MU framework based on embedding techniques to map data into vectors and then be stored in vector databases. Our approach exploits the overlap between CIL and MU tasks for acceleration. Experiments demonstrate the capability of achieving unlearning effectiveness and orders of magnitude (upto ~ 278×) of acceleration.

Task assignment is a fundamental problem in spatial crowdsourcing. In many spatial crowdsourcing platforms, such as Didi, AMAP, and Uber, there are hybrid tasks, including real-time and reservation-type tasks, which are with different constraints and unevenly distributed in spatial and temporal. For these hybrid tasks, most existing studies suffer from low task completion rate and low profit for two reasons: firstly, they focus on homogeneous tasks with uniform constraints, and assign hybrid tasks separately; secondly, they cannot effectively address the uneven distribution of hybrid tasks. Inspired by this, we delve into the problem of online hybrid task assignment (HyTAO) with the goal of maximizing total revenue by simultaneously assigning both real-time and reservation-type tasks online for the first time. We prove the NP-hardness of the offline version of the HyTAO problem. To solve HyTAO effectively, we utilize a cross-platform cooperation model to tackle the challenge of non-uniform distribution. Following this, we design a binary tree-based search algorithm, namely BTS, which is capable of uniformly processing various types of tasks and quickly searching for available workers. Additionally, we discuss the parallel optimization strategies of BTS. To further enhance performance, we develop TBTS, which identifies tasks with high increased revenue based on a threshold. Finally, we conduct a comprehensive analysis of the complexity and competitive ratio of both BTS and TBTS. Extensive experiments are performed to demonstrate the efficiency of our approaches.

Continuous subgraph matching (CSM) is a critical task for analyzing dynamic graphs and has a wide range of applications, such as merchant fraud detection, cyber-attack hunting, and rumor detection. Although many efficient CSM algorithms have been recently proposed, they are mainly designed to process a single query. However, in some application scenarios, multi-query oriented continuous subgraph matching (MQCSM) may be of more interest. To our knowledge, the two existing solutions to MQCSM are outdated due to unsatisfactory performance. In this paper, we propose MQ-Match, an efficient approach to MQCSM. First, we design a compact yet effective index structure CCG, which maintains the local matching result of vertices in the data graph using a directed graph. The directed edges in CCG can be utilized as an effective pruning rule for the subsequent incremental matching algorithm when expanding a partial match. Then, we develop a computation sharing incremental matching algorithm. In specific, a set of matching trees is constructed based on the depth-first search trees of the query graphs. By utilizing CCG, we conduct subgraph matching for the matching tree to collect the incremental matches for the query graphs, where the common structures of query graphs are matched only once. Extensive experiments show that MQ-Match can achieve 3.1x-7071.4x speedup over the competitors, and consumes much less memory under the majority of the experiment settings.

Tensor data is widely used in fields such as smart grids, cloud systems, and deep learning. As the scale of this data increases, storage and transmission costs rise significantly. Many tensor data exhibit low-rank structures, offering the potential for data compression through low-rank decomposition techniques. Tucker decomposition, a typical low-rank decomposition technique, achieves data compression and interpretability by capturing complex data correlations and representing the original tensor with a compact core tensor and factor matrices. However, the compression ratio provided by Tucker decomposition is often insufficient, particularly for large-scale tensors. To tackle this issue, we propose a robust low-rank tensor compression method that leverages Tucker decomposition with quantization and coding. Initially, we establish a robust Tucker decomposition framework that decomposes the low-rank tensor into a core tensor and factor matrices using well-designed Tucker rank-setting rules. This framework effectively handles noise and missing values. Subsequently, we conduct an in-depth analysis of the numerical characteristics of the core tensor and factor matrices within the Tucker decomposition framework. Based on their distinct characteristics, we design tailored quantization and coding schemes to compress the core tensor and factor matrices, respectively, thereby significantly improving the compression ratio of Tucker decomposition while maintaining high accu-racy. Through extensive experiments on four publicly available datasets (which can form 3 or 4 order tensors), we demonstrate that our approach can achieve compression ratios$4\times-10\times$higher than the best competitor, with recovery errors improved by 8% - 42%.

With the popularity of GPS-equipped smart devices, spatial crowdsourcing (SC) techniques have attracted growing attention in both academia and industry. In existing trajectory-aware task assignment approaches, tasks assigned to a worker may be far apart from each other, resulting in a higher detour cost as the worker needs to deviate from the original trajectory more often than necessary. Motivated by the above observations, we investigate a trajectory-aware task coalition assignment (TCA) problem and prove it to be NP-hard. The goal is to maximize the number of assigned tasks by assigning task coalitions to workers based on their preferred trajectories. To tackle the TCA problem, we develop a batch-based three-stage framework consisting of task grouping, planning, and assignment. Extensive experiments on real and synthetic datasets demonstrate the effectiveness and efficiency of the proposed algorithms.

We propose DeformingNet, an effective 3D point cloud completion network. Unlike existing methods that complete partial point cloud by directly learning the morphing function from 2D grids to 3D shapes, which limits the model’s inference capability due to the intrinsic gaps between feature spaces with different dimensions, we design a deforming-based point generator that emulates the deforming from multiple uniform 3D priors (i.e., pre-defined 3D point clouds in cube shape) into 3D shapes. In addition, we design a MAE-based encoder, which introduces the MAE encoder from point-MAE pre-trained on ShapeNet dataset to learn the correlation of local regions and fuse local and global features to enrich the latent representation. The 3D shapes generated by DeformingNet have both accurate local details and faithful global structure with less noise. Experiments demonstrate that our DeformingNet outperforms state-of-the-art methods in terms of quantitative metrics and visual quality.

Although person search has achieved great progress with advanced deep neural networks, it remains an intricate challenge, necessitating a concurrent solution to pedestrian detection and person re-identification. Due to the difficulty of accurate labeling, the current widely used person search datasets contain numerous unknown identities in addition to the carefully labeled known ones. However, previous efforts usually simply utilize these unknown persons as negative samples, inadvertently disregarding the wealth of untapped sample information. To tackle this issue, we propose a novel Unknown Instance Learning (UIL) method that introduces an extra network branch to mine the potential identity information of unknown individuals during training, apart from the fundamental search branch. Specifically, a Clustering-Based Unknowns Labeling (CBUL) mechanism is introduced to effectively exploit the features extracted from unknown persons. In addition, recognizing the distinct roles of pseudo and real identity labels in discriminative feature learning, we design a Reformative Online Instance Matching (ROIM) loss and maintain lookup tables for real and pseudo labels separately. Extensive experiments on the CUHK-SYSU and PRW datasets demonstrate the superiority of our approach in comparison to other state-of-the-art person search models.

Models trained on ultrasound images from one institution typically experience a decline in effectiveness when transferred directly to other institutions. Moreover, unlike natural images, dense and overlapped structures exist in fetus ultrasound images, making the detection of structures more challenging. Thus, to tackle this problem, we propose a new Unsupervised Domain Adaptation (UDA) method named ToMo-UDA for fetus structure detection, which consists of the Topology Knowledge Transfer (TKT) and the Morphology Knowledge Transfer (MKT) module. The TKT leverages prior knowledge of the medical anatomy of fetal as topological information, reconstructing and aligning anatomy features across source and target domains. Then, the MKT formulates a more consistent and independent morphological representation for each substructure of an organ. To evaluate the proposed ToMo-UDA for ultrasound fetal anatomical structure detection, we introduce FUSH$^2$, a new Fetal UltraSound benchmark, comprises Heart and Head images collected from Two health centers, with 16 annotated regions. Our experiments show that utilizing topological and morphological anatomy information in ToMo-UDA can greatly improve organ structure detection. This expands the potential for structure detection tasks in medical image analysis.

With the rapid development of artificial intelligence, a substantial number of computing-intensive applications have emerged in Internet of Things (IoT) devices. The mobile edge computing (MEC) architecture enables the provision of abundant computing and storage resources in close proximity to end users (EUs), thereby effectively enhancing their quality of experience (QoE). Nonetheless, both the MEC server and EUs are self-interests, it is crucial to establish suitable incentive mechanism to promote active engagement from both parties in the offloading process. Therefore, we employ the Stackelberg game to describe the interaction process between EUs and the MEC server, and an optimal relationship between bandwidth and offloading task size is established to simplify the decision problem for EUs. Then, the optimal strategies for the MEC server and EUs are solved using reverse induction. Given the limited resources of the MEC server, we propose a dynamic programming-based resource allocation (DPRA) algorithm to maximize the revenue of the MEC server while ensuring the cost of each EU. The simulation results demonstrate that the DPRA algorithm can reduce latency and energy consumption costs, significantly outperforming other comparative strategies in terms of performance at both EUs and the MEC server.

Graph pattern query is a powerful tool for extracting crucial information from property graphs. With the exponential growth of sizes, property graphs are typically divided into multiple subgraphs (referred to as partitions ) and stored across various sites in distributed environments. Existing graph partitioning methods have not been efficiently optimized for pattern queries, resulting in numerous query matches across multiple partitions, called crossing matches. Identifying these matches requires much inter-partition communication, which is the primary performance bottleneck in distributed query processing. To address this issue, this paper introduces a novel connectivity-oriented relationship-disjoint partitioning method, namely RCP (Relationship Connectivity Partitioning), aimed at enhancing the efficiency of graph pattern query processing by reducing crossing matches. By employing each weakly connected component of the subgraph, which is induced by different relationship labels, as a basic unit of partition, RCP ensures that matches for both variable-length path and labeled graph pattern queries are not crossing matches. Here, variable-length path and labeled graph pattern are two common components in graph pattern queries to identify paths meeting specific label constraints and retrieve subgraphs with consistent relationship types, respectively. Moreover, in the query processing phase, we further demonstrate that all graph pattern queries, belonging to these two basic queries or their extensions, can be executed independently under RCP, thereby avoiding crossing matches. In experiments, we implemented two prototype distributed property graph systems based on Neo4j and JanusGraph, which use declarative and functional query language, respectively. Experimental results on billion-scale datasets demonstrate that our approach brings a performance improvement of nearly two orders of magnitude over state-of-the-art partitioning methods.

The enumeration of hop-constrained simple paths is a building block in many graph-based areas. Due to the enormous search spaces in large-scale graphs, a single machine can hardly satisfy the requirements of both efficiency and memory, which causes an urgent need for efficient distributed methods. In practice, it is inevitable to produce plenty of intermediate results when directly extending centralized methods to the distributed environment, thereby causing a memory crisis and weakening the query performance. The state-of-the-art distributed method HybridEnum designed a hybrid search paradigm to enumerate simple paths. However, it makes massive exploration for the redundant vertices not located in any simple path, thereby resulting in poor query performance. To alleviate this problem, we design a distributed approach DistriEnum to optimize query performance and scalability with well-bound memory consumption. Firstly, DistriEnum adopts a graph reduction strategy to rule out the redundant vertices without satisfying the constraint of hop number. Then, a core search paradigm is designed to simultaneously reduce the traversal of shared subpaths and the storage of intermediate results. Moreover, DistriEnum is equipped with a task division strategy to theoretically achieve workload balance. Finally, a vertex migration strategy is devised to reduce the communication cost during the enumeration. The comprehensive experimental results on 10 real-world graphs demonstrate that DistriEnum achieves up to 3 orders of magnitude speedup than HybridEnum in query performance and exhibits superior performances on scalability, communication cost, and memory consumption.

GPUs are commonly utilized to accelerate GNN training, particularly on a multi-GPU server with high-speed interconnects (e.g., NVLink and NVSwitch). However, the rapidly increasing scale of graphs poses a challenge to applying GNN to real-world applications, due to limited GPU memory. This paper presents XGNN, a multi-GPU GNN training system that fully utilizes system memory (e.g., GPU and host memory), as well as high-speed interconnects. The core design of XGNN is the Global GNN Memory Store (GGMS), which abstracts underlying resources to provide a unified memory store for GNN training. It partitions hybrid input data, including graph topological and feature data, across both GPU and host memory. GGMS also provides easy-to-use APIs for GNN applications to access data transparently, forwarding data access requests to the actual physical data partitions automatically. Evaluation on various multi-GPU platforms using three common GNN models with four large-scale datasets shows that XGNN outperforms DGL, Quiver and DGL+C by up to 7.9X (from 2.3X), 15.7X (from 3.3X) and 2.8X (from 1.3X), respectively.

The legal judgment prediction (LJP) task is to automatically decide appropriate law articles, charges, and term of penalty for giving the fact description of a law case. It considerably influences many real legal applications and has thus attracted the attention of legal practitioners and AI researchers in recent years. In real scenarios, many confusing charges are encountered, which makes LJP challenging. Intuitively, for a controversial legal case, legal practitioners usually first obtain various possible judgment results as candidates based on the fact description of the case; then these candidates generally need to be carefully considered based on the facts and the rationality of the candidates. Inspired by this observation, this paper presents a novelRecall &Ranking framework, dubbed as$\boldsymbol{R}^{2}$, which attempts to formalize LJP as a two-stage problem. The recall stage is designed to collect high-likelihood judgment results for a given case; these results are regarded as candidates for the ranking stage. The ranking stage introduces a verification technique to learn the relationships between the fact description and the candidates. It treats the partially correct candidates as semi-negative samples, and thus has a certain ability to distinguish confusing candidates. Moreover, we devise a comprehensive judgment strategy to refine the final judgment results by comprehensively considering the rationality of multiple probable candidates. We carry out numerous experiments on two widely used benchmark datasets. The experimental results demonstrate our proposed approach's effectiveness compared to the other competitive baselines.

Federated learning (FL) enables a great deal of distributed independent participants to collaborate in training without sharing data. Malicious adversary can poison the local model by backdoor poisoning attacks and utilize the characteristic that server cannot trace original data to make the poisoned model directly aggregated. Especially in the AIoT-FL network that generates large amounts of data in real-time, such an attack is more powerful. In this paper, we design a sybil-based backdoor poisoning attacks (SBPA) against the above vulnerability. Malicious participants inject backdoor triggers into distributed Big Data to covertly complete data poisoning. During subsequent iterative aggregation, the joint model activates the backdoor during testing to achieve misclassification for the backdoor images. Besides, malicious participants create some sybil nodes to join the aggregation by taking advantage of the vulnerability of system devices to be easily disconnected. They are committed to making the poisoned local models aggregated with higher probability. Their goal works on making the final global model misclassify backdoor images while keeping high classification accuracy on the other non-backdoor samples. We conduct extensive experiments on multiple datasets and exhibit more robust performance than the state-of-the-art in various metrics under both data distribution including i.i.d. and non-i.i.d. scenarios.

Experienced developers often leverage well-tuned libraries and allocate their routines for computing tasks to enhance performance when building modern scientific and engineering applications. However, such well-tuned libraries are meticulously customized for specific target architectures or environments. Additionally, the performance of their routines is significantly impacted by the actual input data of computing tasks, which often remains uncertain until runtime. Accordingly, statically allocating these library routines may hinder the adaptability of applications and compromise performance, particularly in the context of heterogeneous systems. To address this issue, we propose the Compiler-Assisted Adaptive Library Routines Allocation (COALA) framework for heterogeneous systems. COALA is a fully automated mechanism that employs compiler assistance for dynamic allocation of the most suitable routine to each computing task on heterogeneous systems. It allows the deployment of varying allocation policies tailored to specific optimization targets. During the application compilation process, COALA reconstructs computing tasks and inserts a probe for each of these tasks. Probes serve the purpose of conveying vital information about the requirements of each task, including its computing objective, data size, and computing flops, to a user-level allocation component at runtime. Subsequently, the allocation component utilizes the probe information along with the allocation policy to assign the most optimal library routine for executing the computing tasks. In our prototype, we further introduce and deploy a performance-oriented allocation policy founded on a machine learning-based performance evaluation method for library routines. Experimental verification and evaluation on two heterogeneous systems reveal that COALA can significantly improve application performance, with gains of up to 4.3x for numerical simulation software and 4.2x for machine learning applications, and enhance system utilization by up to 27.8%.

Cache miss analysis has become one of the most important things to improve the execution performance of a program. Generally, the approaches for analyzing cache misses can be categorized into dynamic analysis and static analysis. The former collects sampling statistics during program execution but is limited to specialized hardware support and incurs expensive execution overhead. The latter avoids the limitations but faces two challenges: inaccurate execution path prediction and inefficient analysis resulted by the explosion of the program state graph. To overcome these challenges, we propose CBANA, an LLVM- and process address space-based lightweight, efficient, and flexible cache behavior analysis framework. CBANA significantly improves the prediction accuracy of the execution path with awareness of inputs. To improve analysis efficiency and utilize the program preprocessing, CBANA refactors loop structures to reduce search space and dynamically splices intermediate results to reduce unnecessary or redundant computations. CBANA also supports configurable hardware parameter settings, and decouples the module of cache replacement policy from other modules. Thus, its flexibility is established. We evaluate CBANA by using the popular open benchmark PolyBench, graph workloads, and our synthetic workloads with good and poor data locality. Compared with the popular dynamic cache analysis tools Perf and Valgrind, the cache miss gap is less than 3.79% and 2.74% respectively with over ten thousand data accesses for the synthetic workloads, and the time reduction is up to 92.38% and 97.51% for the multiple-path workloads. Compared with the popular static cache analysis tool Heptane, CBANA achieves a time reduction of 97.71% while ensuring accuracy at the same time.

Remarkable breakthroughs but daunting complexities of deep learning have aroused widespread interest in dedicated deep neural network (DNN) acceleration hardware, among which optical accelerators (OAs) are particularly promising thanks to their unprecedentedly high-performance-per-watt. However, the development of OAs is much slower than that of electrical accelerators due to threefold challenges. First, the OA design space is ample and discrete, making it tough for OA optimization; Second, the ecosystem that facilitates OA development is still in its infancy. Techniques to support OA design remain less explored, limiting both the achievable performance and the innovative development of OAs; and Third, OAs are highly sensitive to fabrication-induced process variations and thermal fluctuations (i.e., PTVs), which degrades OAs’ inference robustness and even renders them unusable in practice. In this article, we develop AutOAS, the first-of-its-kind framework for Automated Optical Accelerator Search, in order to jointly boost acceleration efficiency, inference robustness, and development speed. Our AutOAS comprises four enabling components: 1) a holistic OA search space, which takes full consideration of OAs’ micro-architectures (e.g., the type, shape and size of core functional units for data computation and data access), dataflow choices, DNN-to-accelerator mapping methods, memory hierarchy and PTV mitigation techniques; 2) a PTV Regulator, which can emulate the impact of PTVs on OAs’ inference accuracy based on given PTV profiles, and enables energy-efficient PTV mitigation on OAs; 3) an O-Performance Predictor, which enables accurate yet efficient predictions of an OA’s energy, throughput (latency) and chip area according to the DNN model and OA architecture parameters; and 4) two O-Search Engines (i.e., a differentiable search engine and an evolutionary search engine), which can automatically explore the large design space of OAs and identify the optimal accelerators to maximize the acceleration targets. Based on 10 DNN models widely applied in both computer vision and sequence modeling tasks, extensive experiments and ablation studies validate the effectiveness of our PTV Regulator, O-Performance Predictor, and O-Search Engines, as well as the superior performance of AutOAS-generated OAs.

As promising industrial embedded networks, Controller Area Networks with Flexible Data-rate (CAN-FD) are widely used in time-sensitive domains, such as automotive networks. However, the absence of built-in security mechanisms in CAN-FD necessitates the development of security protection mechanisms. The existing Lightweight Authentication for Secure Automotive Networks (LASAN) framework focuses on enhancing the security of CAN/CAN-FD communication but neglects the conflict between security and delay. In this study, we conduct a thorough analysis of the causal mechanism related to the security and delay of LASAN and propose a static message scheduling method called Cooperative Swapping Approach (CSA) to achieve secure and low-delay CAN-FD communication. CSA is to minimize the end-to-end delay of precedence-constrained CAN-FD applications by swapping message positions in a valid message sequence. Nevertheless, exchanging message positions may impact the precedence dependencies between messages; therefore, we propose a novel Cooperative Transform Approach (CTA) within the CSA to efficiently preserve these precedence constraints. Valid message sequences with minimal end-to-end delays of a motivation example and an Adaptive Cruise Control (ACC) application are obtained in LASAN by CSA. These sequences are implemented on the embedded microcontroller platform of STM32H743IITs for evaluation. Experimental results show that our proposed CSA can effectively reduce the end-to-end delay of LASAN and outperform the state-of-the-art static message scheduling method in terms of low delay.

3-D placement is an important stage in 3-D physical synthesis. In addition to the need to place the standard cells or Macros inside the die, the placement of interdie connections also needs to be considered. As the density of the interdie connections gradually increases, their placement becomes more critical. However, unlike standard cells, the interdie connection does not need to be aligned into rows, which leads to a larger legalization solution space. Legalization aims at minimizing the total and maximum displacement to maintain the quality of global placement on the premise of no violation of the physical circuit constraint. In this article, we propose a two-stage legalization approach (NGLIC) for nonaligned-row interdie connections. In the initial stage, a single-row height legalization algorithm is used for dense placement. Afterward, the unequal multirow height legalization (UML) is designed for sparse placement in the post-optimization stage. A pruning scheme is adopted to identify and eliminate redundant computations. The performance, effectiveness, and configuration are analyzed empirically based on ICCAD 2022 benchmarks. Compared to the state-of-the-art multirow or single-row height legalization, our approach outperforms well-known algorithms, such as multirow global legalization (MGL) and Abacus by at least 24% averaged total displacement, 3% averaged maximum displacement, and 31% averaged HPWL Growth. Our case study also illustrates the effectiveness of UML and NGLIC. In addition, the effectiveness of pruning is also validated by experiments that show savings of 33% redundant computations.

Echocardiography is an essential procedure for the prenatal examination of the fetus for congenital heart disease (CHD). Accurate segmentation of key anatomical structures in a four-chamber view is an essential step in measuring fetal growth parameters and diagnosing CHD. Currently, most obstetricians perform segmentation tasks manually, but the pixel-level operation is labor-intensive and requires extensive anatomical knowledge and clinical experience. As such, efficiently and accurately detecting structures from real-world fetal ultrasound images is a key challenge. In this paper, we propose a YOLOX-based deep instance segmentation neural network (i.e., IS-YOLOX) for cardiac anatomical structure location and segmentation in fetal ultrasound images. Specifically, we reconstruct a new instance segmentation branch based on a multi-task deep learning framework. We then design a new multi-level non-maximum suppression (NMS) mechanism to further improve the segmentation performance that consists of three levels of selection. Moreover, unlike two-stage instance segmentation approaches, our method does not rely on object detection results. To the best of our knowledge, this is the first study regarding instance segmentation on 13 types of anatomical structures in the fetal four-chamber view. Extensive experiments were carried out on clinical datasets, and the experimental results show that our method outperforms nine competitive baselines.

Large-scale cloud applications today are often deployed using multiple containers, and a container overlay network is the de facto method to provide connectivity among these containers. However, the existing tunneling-based overlay network incurs significant performance overhead due to the need of transformation for every packet. Recent workSlim, through manipulating connection-level meta-data, allows containers to use host OS sockets directly thus they can achieve good performance without extra packet tunneling. Nevertheless, the connection setup is significantly slowed down, which requires an extra round-trip communication between both sides to pass the mapping information of the host OS socket and the container socket. This greatly hurts the performance of many cloud applications that must process short connections at high speed. We proposeSlimFast, a low-overhead container overlay network which provides a fast connection setup.SlimFastdirectly uses the host OS socket for container communication asSlim. However,SlimFastneeds no extra communication during connection setup. We reserve a dedicated host port for the container network and use socket mapping table to locally find the right container socket during connection setup. We implementSlimFastwhich is compatible with existing container applications. Experiments show that,SlimFastcan improve the connection setup time by about 2.1x compared withSlim, meanwhile maintaining low-overhead during data transmission asSlim. This brings significant performance improvement to real applications. Particularly, testbed results show thatSlimFastimproves the throughput of Nginx proxy and Memcached by about 0.9x and 2.2x, respectively.

With the increasing scale of cloud computing applications of next-generation embedded systems, a major challenge that domain scientists are facing is how to efficiently store and analyze the vast volume of output data. Compression can reduce the amount of data that needs to be transferred and stored. However, most of the large datasets are in floating-point format, which exhibits high entropy. As a result, existing lossless compressors cannot provide enough performance for such applications. To address this problem, we propose a total variation reduction method for improving the compression ratio of lossless compressors (namely, FPC + and FPZIP + ), which employs a median-based hyperplane to precondition the data. In particular, we first try to exploit the space-filling curve (SFC), a well-known technique to preserve data locality for a multi-dimensional dataset. We show and explain why a raw SFC, such as Hilbert and Z-order curves, cannot improve the compression ratio. Then, we explore the opportunity and theoretical feasibility of the proposed total variation reduction-based algorithm. The experiment results show the effectiveness of the proposed method. The compression ratios are improved up to 48.2% (20.6% on average) for FPZIP and 42.4% (18.4% on average) for FPC. Moreover, through observing the time composition of the proposed method, it is found that the median finding holds a high percentage of the execution time. Hence, we further introduce an approximate median finding algorithm, providing a linear-time overhead reduction scheme. The experiment results clearly demonstrate that this algorithm reduces execution time by an average of 56.7% and 40.7% compared to FPC + and FPZIP + , respectively.

One area of current research on adversarial attacks is how to generate plausible adversarial examples when only a small number of datasets are available. Current adversarial attack algorithms used to attack these black-box systems face a number of challenges, such as difficulty in training convergence, ambiguous sample images, substitute models collapse, unsatisfactory attack success rates, high query cost, and low defense capability improvement of target models. As a result, constructing plausible adversarial situations in a few known real-world sample circumstances remains difficult. As a solution to the aforementioned issues, this study introduces MC-Net, a novel multi-stage and multi-class balanced generating method based on a limited number of samples to generate realistic adversarial examples. Firstly, a multi-task learning approach is used to train the GAN by fully utilizing the small samples, ensuring that the size of the generated dataset for each category is balanced. In addition, we design a weight-balancing strategy to ensure faster convergence of each sub-network. Then, in the second stage, the generated samples of different categories are used to train a substitute model, and the distillation method is adopted to learn the output distribution of the target model. Finally, adversarial examples are constructed on the generated samples to complete the attack on the target models. Extensive experiments have proven that MC-Net has the following advantages: 1) The substitute model converges quickly using limited samples and queries; 2) High attack success rates can be obtained with a few queries; and 3) The constructed adversarial examples significantly improve the target model’s defense. Furthermore, we only utilize a few queries for the Microsoft Azure online model to obtain a satisfactory result. Our code can be found at https://github.com/jiaokailun/A-fast.

Adversarial attack algorithms are useful for testing and improving the robustness of industrial AI models. However, attacking black-box models with limited queries and unknown real labels remains a significant challenge. To overcome this challenge, we propose using contrastive learning to train a generated substitute model called attack contrastive learning network (ACL-Net) to attack black-box models with very few queries and no real labels. ACL-Net achieves end-to-end contrastive learning during training without labels, which differs from previous contrastive learning methods that required separate training for the classification layer with labels. We improve ACL-Net's robustness by using adversarial examples to train it during the attack stage. This approach results in more effective adversarial examples generated by ACL-Net. We conducted extensive experiments to validate the effectiveness of ACL-Net. Compared with the latest algorithms, ACL-Net requires fewer queries to achieve better attack performance, demonstrating its superiority in query-efficient black-box attacks. Overall, our approach presents a promising solution to the challenge of attacking black-box models with limited queries and unknown real labels. Our results show the effectiveness of using contrastive learning to train generated substitute models, and the potential for improving the robustness of industrial AI models through adversarial attacks.

Monocular depth estimation has made considerable advances under clear weather conditions. However, how to learn accurate scene depth under rain and fog conditions and alleviate the negative influence of occlusion, light, visibility, etc., is an open problem. To address this problem, in this article, we split the adverse weather depth estimation network into two subbranches: the depth prediction branch and the masked encoding branch. The depth prediction branch is used for depth estimation. The masked encoding branch, inspired by masked image modeling, uses random masks to simulate occlusion or low visibility often seen in rain and fog, forcing this branch to learn to infer the prediction of masked regions from the context. In order to make the masked encoding better enhance the depth prediction, we designed the mask feature fusion module, which can fuse the depth and spatial context features of the two branches to produce a fine-level depth map. The experimental results on the Foggy Cityscapes and RainCityscapes datasets demonstrate that our method achieves state-of-the-art performance, significantly outperforming previous methods across all evaluation metrics.

Congenital heart disease (CHD) is the most common congenital disability affecting healthy development and growth, even resulting in pregnancy termination or fetal death. Recently, deep learning techniques have made remarkable progress to assist in diagnosing CHD. One very popular method is directly classifying fetal ultrasound images, recognized as abnormal and normal, which tends to focus more on global features and neglects semantic knowledge of anatomical structures. The other approach is segmentation-based diagnosis, which requires a large number of pixel-level annotation masks for training. However, the detailed pixel-level segmentation annotation is costly or even unavailable. Based on the above analysis, we propose SKGC, a universal framework to identify normal or abnormal four-chamber heart (4CH) images, guided by a few annotation masks, while improving accuracy remarkably. SKGC consists of a semantic-level knowledge extraction module (SKEM), a multi-knowledge fusion module (MFM), and a classification module (CM). SKEM is responsible for obtaining high-level semantic knowledge, serving as an abstract representation of the anatomical structures that obstetricians focus on. MFM is a lightweight but efficient module that fuses semantic-level knowledge with the original specific knowledge in ultrasound images. CM classifies the fused knowledge and can be replaced by any advanced classifier. Moreover, we design a new loss function that enhances the constraint between the foreground and background predictions, improving the quality of the semantic-level knowledge. Experimental results on the collected real-world NA-4CH and the publicly FEST datasets show that SKGC achieves impressive performance with the best accuracy of 99.68% and 95.40%, respectively. Notably, the accuracy improves from 74.68% to 88.14% using only 10 labeled masks.

In the fetal cardiac ultrasound examination, standard cardiac cycle (SCC) recognition is the essential foundation for diagnosing congenital heart disease. Previous studies have mostly focused on the detection of adult CCs, which may not be applicable to the fetus. In clinical practice, localization of SCCs needs to recognize end-systole (ES) and end-diastole (ED) frames accurately, ensuring that every frame in the cycle is a standard view. Most existing methods are not based on the detection of key anatomical structures, which may not recognize irrelevant views and background frames, results containing non-standard frames, or even it does not work in clinical practice. We propose an end-to-end hybrid neural network based on an object detector to detect SCCs from fetal ultrasound videos efficiently, which consists of 3 modules, namely Anatomical Structure Detection (ASD), Cardiac Cycle Localization (CCL), and Standard Plane Recognition (SPR). Specifically, ASD uses an object detector to identify 9 key anatomical structures, 3 cardiac motion phases, and the corresponding confidence scores from fetal ultrasound videos. On this basis, we propose a joint probability method in the CCL to learn the cardiac motion cycle based on the 3 cardiac motion phases. In SPR, to reduce the impact of structure detection errors on the accuracy of the standard plane recognition, we use XGBoost algorithm to learn the relation knowledge of the detected anatomical structures. We evaluate our method on the test fetal ultrasound video datasets and clinical examination cases and achieve remarkable results. This study may pave the way for clinical practices.

Biometric parameter measurements are powerful tools for evaluating a fetus's gestational age, growth pattern, and abnormalities in a 2D ultrasound. However, it is still challenging to measure fetal biometric parameters automatically due to the indiscriminate confusing factors, limited foreground-background contrast, variety of fetal anatomy shapes at different gestational ages, and blurry anatomical boundaries in ultrasound images. The performance of a standard CNN architecture is limited for these tasks due to the restricted receptive field. We propose a novel hybrid Transformer framework, TransFSM, to address fetal multi-anatomy segmentation and biometric measurement tasks. Unlike the vanilla Transformer based on a single-scale input, TransFSM has a deformable self-attention mechanism so it can effectively process multi-scale information to segment fetal anatomy with irregular shapes and different sizes. We devised a BAD to capture more intrinsic local details using boundary-wise prior knowledge, which compensates for the defects of the Transformer in extracting local features. In addition, a Transformer auxiliary segment head is designed to improve mask prediction by learning the semantic correspondence of the same pixel categories and feature discriminability among different pixel categories. Extensive experiments were conducted on clinical cases and benchmark datasets for anatomy segmentation and biometric measurement tasks. The experiment results indicate that our method achieves state-of-the-art performance in seven evaluation metrics compared with CNN-based, Transformer-based, and hybrid approaches. By Knowledge distillation, the proposed TransFSM can create a more compact and efficient model with high deploying potential in resource-constrained scenarios. Our study serves as a unified framework for biometric estimation across multiple anatomical regions to monitor fetal growth in clinical practice.

Accurate recognition of fetal anatomical structure is a pivotal task in ultrasound (US) image analysis. Sonographers naturally apply anatomical knowledge and clinical expertise to recognizing key anatomical structures in complex US images. However, mainstream object detection approaches usually treat each structure recognition separately, overlooking anatomical correlations between different structures in fetal US planes. In this work, we propose a Fetal Anatomy Reasoning Network (FARN) that incorporates two kinds of relationship forms: a global context semantic block summarized with visual similarity and a local topology relationship block depicting structural pair constraints. Specifically, by designing the Adaptive Relation Graph Reasoning (ARGR) module, anatomical structures are treated as nodes, with two kinds of relationships between nodes modeled as edges. The flexibility of the model is enhanced by constructing the adaptive relationship graph in a data-driven way, enabling adaptation to various data samples without the need for predefined additional constraints. The feature representation is further refined by aggregating the outputs of the ARGR module. Comprehensive experimental results demonstrate that FARN achieves promising performance in detecting 37 anatomical structures across key US planes in tertiary obstetric screening. FARN effectively utilizes key relationships to improve detection performance, demonstrates robustness to small-scale, similar, and indistinct structures, and avoids some detection errors that deviate from anatomical norms. Overall, our study serves as a resource for developing efficient and concise approaches to model inter-anatomy relationships.

How to construct imperceptible (realistic) fake samples is critical in adversarial attacks. Due to the sample feature diversity of a recommender system (containing both discrete and continuous features), traditional gradient-based adversarial attack methods may fail to construct realistic fake samples. Meanwhile, most recommendation models adopt click-through rate (CTR) predictors, which usually utilize black-box deep models with discrete features as input. Thus, how to efficiently construct realistic fake samples for black-box recommender systems is still full of challenges. In this article, we propose a hierarchical adversarial attack method against black-box CTR models via generating realistic fake samples, named CTRAttack. To better train the generation network, the weights of its embedding layer are shared with those of the substitute model, with both the similarity loss and classification loss used to update the generation network. To ensure that the discrete features of the generated fake samples are all real, we first adopt the similarity loss to ensure that the distribution of the generated perturbed samples is sufficiently close to the distribution of the real features, and then the nearest neighbor algorithm is used to retrieve the most appropriate features for non-existent discrete features from the candidate instance set. Extensive experiments demonstrate that CTRAttack can not only effectively attack the black-box recommender systems but also improve the robustness of these models while maintaining prediction accuracy.

The meteorological environment has a profound impact on global health (e.g., air quality), science and technology (e.g., rocket launches), and economic development (e.g., poverty reduction) etc. Meteorological prediction presents numerous challenges to both academia and industry due to its multifaceted nature which encompasses real-time observations and complex modeling. Recent research adopt graph convolutional recurrent network and establish coordinate information to obtain local spatial-temporal pattern. However, the model only utilizes the local spatial-temporal information and fail to fully consider the dynamic meteorological situation. To address the above limitations, we propose a Dynamic Spatial-Temporal Graph Neural Network (DSTGNN) to learn global-local meteorological features. Specifically, we divide the global spatial-temporal information along the timeline to obtain local spatial-temporal information. For the global aspect, we design a random throwedge module during the neighborhood propagation process in graph neural network (GNN) to extract the features and adapt to the dynamic situation. We also establish convolution operation module to learn the features. Next, we perform information fusion on the two modules to capture sufficient features. In addition, we employ graph ordinary differential equation (ODE) network and utilize the coordinate information to obtain the long-term features and coordinate relationships. In the local aspect, we first construct a GNN to conduct graph embedding. Then, we integrate another GNN into a gated recurrent unit (GRU) and also use the coordinate information to explore the features and coordinate relationships. Finally, we combine the global and local features via a global-local features learning layer for meteorological prediction. Experimental results on the four real-world meteorological datasets show that DSTGNN outperforms the baseline models.

Unsupervised multi-view bipartite graph clustering (MVBGC) is a fast-growing research, due to promising scalability in large-scale tasks. Although many variants are proposed by various strategies, a common design is to construct the bipartite graph directly from the input data, i.e. only consider the unidirectional “encoding” process. However, “encoding-decoding” mechanism is a popular design for deep learning, the most representative one is auto-encoder (AE). Enlightened by this, this paper rethinks existing MVBGC paradigms and transfers the “encoding-decoding” design into graph machine learning, and proposes a novel framework termed auto-encoding multi-view bipartite graph clustering (BGAE), which integrates encoding, bipartite graph construction, and decoding modules in a self-supervised learning manner. The encoding module extracts a latent joint representation from the input data, the bipartite graph construction module learns a bipartite graph with connectivity constraint in latent semantic space, and the decoding module recreates the input data via the bipartite graph. Therefore, our novel BGAE combines representation learning, bipartite graph learning, reconstruction learning, and label inference into a unified framework. All the modules are seamlessly integrated and mutually reinforcing for clustering-friendly purposes. Extensive experiments verify the superiority of our novel design and the significance of “decoding” process. To the best of our knowledge, this is the first attempt to explore “encoding-decoding” design in traditional MVBGC. The code is provided athttps://github.com/liliangnudt/BGAE.

In recommendation systems, Graph Convolutional Networks (GCNs) often suffer from significant computational and memory cost when propagating features across the entire user-item graph. While various sampling strategies have been introduced to reduce the cost, the challenge of neighbor explosion persists, primarily due to the iterative nature of neighbor aggregation. This work focuses on exploring subgraph propagation for scalable recommendation by addressing two primary challenges:efficient and effective subgraph constructionandsubgraph sparsity. To address these challenges, we propose a novelGCNmodel for recommendation based onSubgraph propagation, called SubGCN. One key component of SubGCN is BiPPR, a technique that fuses both source- and target-based Personalized PageRank (PPR) approximations, to overcome the challenge ofefficient and effective subgraph construction. Furthermore, we propose a source-target contrastive learning scheme to mitigate the impact ofsubgraph sparsityfor SubGCN. We conduct extensive experiments on two large and two medium-sized datasets to evaluate the scalability, efficiency, and effectiveness of SubGCN. On medium-sized datasets, compared to full-graph GCNs, SubGCN achieves competitive accuracy while using only 23.79% training time on Gowalla and 16.3% on Yelp2018. On large datasets, where full-graph GCNs ran out of the GPU memory, our proposed SubGCN outperforms widely used sampling strategies in terms of training efficiency and recommendation accuracy.

Robust PCA is a popular anomaly detection technique and has been widely used in many applications. Although Robust PCA is promising, it is usually designed in a two-order matrix form, which is inferior to the tensor that can capture multilinearity features of data. Moreover, the detection accuracy under Robust PCA further suffers due to its sensitivity to the rank parameter which is hard to set in practice and the limitation of PCA method in capturing the non-linear feature in the data. To address the issues, we propose a Robust Tensor Convolutional Autoencoder (RTCAE) where the autoencoder instead of SVD is exploited to recover the normal data from the corrupted measurement tensor data. However, directly exploiting deep autoencoder may suffer from the problem of high memory consumption and computation overhead due to the large number of parameters used in autoencoder. To make our anomaly detection lightweight, we further design a Light Convolutional Autoencoder (LightCAE) which contains a compressed autoencoder by exploiting tensor factorization to largely compress the parameters while significantly reducing the computation complexity. We conduct extensive experiments on three real data traces to compare the performance of our proposed schemes (RTCAE and lightCAE) with that of seven baseline algorithms. The experiment results demonstrate that our proposed RTCAE achieves the highest anomaly detection accuracy. Moreover, our LightCAE requires over 60 times smaller memory storage than that required in RTCAE while achieving the similar anomaly detection accuracy.

With the popularity of GPS-equipped smart devices, spatial crowdsourcing (SC) techniques have attracted growing attention in both academia and industry. A fundamental problem in SC is assigning location-based tasks to workers under spatial-temporal constraints. In many real-life applications, workers choose tasks on the basis of their preferred trajectories. However, by existing trajectory-aware task assignment approaches, tasks assigned to a worker may be far apart from each other, resulting in a higher detour cost as the worker needs to deviate from the original trajectory more often than necessary. Motivated by the above observations, we investigate a trajectory-aware task coalition assignment (TCA) problem and prove it to be NP-hard. The goal is to maximize the number of assigned tasks by assigning task coalitions to workers based on their preferred trajectories. For tackling the TCA problem, we develop a batch-based three-stage framework consisting of task grouping, planning, and assignment. First, we design greedy and spanning grouping approaches to generate task coalitions. Second, to gain candidate task coalitions for each worker efficiently, we design task-based and trajectory-based pruning strategies to reduce the search space. Furthermore, a 2-approximate algorithm, termed MST-Euler, is proposed to obtain a route among each worker and task coalition with a minimal detour cost. Third, the MST-Euler Greedy (MEG) algorithm is presented to compute an assignment that results in the maximal number of tasks assigned and a parallel strategy is introduced to boost its efficiency. Extensive experiments on real and synthetic datasets demonstrate the effectiveness and efficiency of the proposed algorithms.

Federated Learning (FL) combined with Differential Privacy (DP) is widespread in healthcare, finance, and IoT due to its advantages in multi-client data distribution. However, existing FL approaches overlook the differential impact levels among clients and data redundancy issues, resulting in high computational overhead and limited real-time applicability. Additionally, non-independent identical distribution (Non-IID) and imbalanced datasets in multi-clients pose challenges in privacy preservation and model overfitting. Therefore, we propose a balanced weight strategy for multi-stage federated learning against multi-client data skewing, called BM-FL, which involves clients, intermediate trust servers (ITSs), and the central server (CS). Firstly, to protect data privacy, an improved Laplace$\epsilon$-differential privacy method is employed. Secondly, a novel generative adversarial network (GAN) called BC-GAN is introduced. It is used to generate realistic fake samples and maintain a balanced proportion of samples across different categories. Then, to make full use of each client's valuable data, we designe a balanced weight strategy. Moreover, extensive experimental results clearly demonstrate the effectiveness of BM-FL in efficiently handling classification tasks involving Non-IID and imbalanced datasets while maintaining privacy and security. Furthermore, our method attains superior classification accuracy with fewer training epochs compared to relevant classical algorithms. The code is available athttps://github.com/ylxzjy/BMFL.git.

Recent research has revealed that deep generative models including flow-based models and Variational Autoencoders may assign higher likelihoods to out-of-distribution (OOD) data than in-distribution (ID) data. However, we cannot sample OOD data from the model. This counterintuitive phenomenon has not been satisfactorily explained and brings obstacles to OOD detection with flow-based models. In this article, we prove theorems to investigate the Kullback-Leibler divergence in flow-based model and give two explanations for the above phenomenon. Based on our theoretical analysis, we propose a new method KLODS to leverage KL divergence and local pixel dependence of representations to perform anomaly detection. Experimental results on prevalent benchmarks demonstrate the effectiveness and robustness of our method. For group anomaly detection, our method achieves 98.1% AUROC on average with a small batch size of 5. On the contrary, the baseline typicality test-based method only achieves 64.6% AUROC on average due to its failure on challenging problems. Our method also outperforms the state-of-the-art method by 9.1% AUROC. For point-wise anomaly detection, our method achieves 90.7% AUROC on average and outperforms the baseline by 5.2% AUROC. Besides, our method has the least notable failures and is the most robust one.

Needle insertion is a critical component of many remote surgical procedures, including biopsies, injections, neurosurgery, and brachytherapy cancer treatments. However, precise visualization of the biopsy needle trajectory remains challenging due to specular reflection, speckle noise, and needle-like anatomical features. This paper proposes a visual feedback prediction framework for ultrasound-assisted percutaneous kidney biopsy in 5G remote surgery, aiming to enhance operator confidence, reduce procedure time, and minimize the risk of unintended bleeding. Building upon this framework, we design a Lightweight-Accuracy Needle Trajectory Prediction (LA-NTP) model by minimizing the backbone and optimizing the multi-module prediction process, incorporating innovative training strategies (i.e., angle-aware geometric and trajectory augmentation losses). The experimental results demonstrate that it achieves competitive performance with only 20.3% of the model size of the previous best real-time method and a 3.7-fold increase in inference speed. Even in challenging scenarios involving large insertion depths and steep angles, our method provides stable and precise navigation.

The construction of undetectable adversarial examples with few perturbances remains a difficult problem in adversarial attacks. At present, most solutions use the standard gradient optimization algorithm to build adversarial examples by applying global perturbations to benign samples and then launch attacks on the targets (e.g., face recognition systems). However, when the perturbance size is limited, the performance of these approaches suffers substantially. The content of crucial places in an image, on the other hand, will impact the final prediction; if these areas can be investigated and limited perturbances introduced, an acceptable adversarial example will be constructed. Based on the foregoing research, this article offers a dual attention adversarial network (DAAN) to produce adversarial examples with limited perturbations. DAAN initially searches for effective areas in an input image using the spatial attention network and channel attention network, and then creates space and channel weights. Following that, these weights direct an encoder and a decoder to generate effective perturbation, which is then combined with the input to produce an adversarial example. Finally, the discriminator determines if the created adversarial examples are true or false, and the attacked model is utilized to determine whether the generated samples fit the attack targets. Extensive studies on various datasets show that DAAN not only delivers the best attack performance across all comparison algorithms with few perturbations, but it can also significantly improve the defensiveness of the attacked models.

Multiple kernel clustering (MKC) is committed to achieving optimal information fusion from a set of base kernels. Constructing precise and local kernel matrices is proven to be of vital significance in applications since the unreliable distant-distance similarity estimation would degrade clustering performance. Although existing localized MKC algorithms exhibit improved performance compared with globally designed competitors, most of them widely adopt the KNN mechanism to localize kernel matrix by accounting for τ -nearest neighbors. However, such a coarse manner follows an unreasonable strategy that the ranking importance of different neighbors is equal, which is impractical in applications. To alleviate such problems, this article proposes a novel local sample-weighted MKC (LSWMKC) model. We first construct a consensus discriminative affinity graph in kernel space, revealing the latent local structures. Furthermore, an optimal neighborhood kernel for the learned affinity graph is output with naturally sparse property and clear block diagonal structure. Moreover, LSWMKC implicitly optimizes adaptive weights on different neighbors with corresponding samples. Experimental results demonstrate that our LSWMKC possesses better local manifold representation and outperforms existing kernel or graph-based clustering algorithms. The source code of LSWMKC can be publicly accessed from https://github.com/liliangnudt/LSWMKC.

Sparse Matrix-Vector Multiplication (SpMV) on GPUs has gained significant attention because of SpMV's importance in modern applications and the increasing computing power of GPUs in the last decade. Previous studies have emphasized the importance of data loading for the overall performance of SpMV and demonstrated the efficacy of coalesced memory access in enhancing data loading efficiency. However, existing approaches fall far short of reaching the full potential of data loading on modern GPUs. In this paper, we propose an efficient algorithm called FastLoad, that speeds up the loading of both sparse matrices and input vectors of SpMV on modern GPUs. Leveraging coalesced memory access, FastLoad achieves high loading efficiency and load balance by sorting both the columns of the sparse matrix and elements of the input vector based on the number of non-zero elements while organizing non-zero elements in blocks to avoid thread divergence. FastLoad takes the Compressed Sparse Column (CSC) format as an implementation case to prove the concept and gain insights. We conduct a comprehensive comparison of FastLoad with the CSC-based SpMV, cuSPARSE, CSR5, and TileSpMV, using the full SuiteSparse Matrix Collection as workload. The experimental results on RTX 3090 Ti demonstrate that our method outperforms the others in most matrices, with geometric speedup means over CSC-based, cuSPARSE, CSR5, and TileSpMV being 2.12×, 2.98×, 2.88×, and 1.22×, respectively.

Sparse tensor contraction (SpTC) is an important operator in tensor networks, which tends to generate a large amount of sparse high-dimensional data, placing higher demands on the computational performance and storage bandwidth of the processor. Using GPUs with powerful arithmetic characteristics is a reliable choice for accelerating SpTC, however, the high dimensionality and sparsity of tensor makes GPU-accelerated SpTC operators suffer from the difficulties of low computational intensity and high memory consumption. The recent introduction of Tensor Core Units (TCUs) on GPUs brings even more powerful arithmetic, which exacerbates the memory wall problem. To cope with the challenges, this paper proposes a new BCB format that linearizes the indices of multidimensional blocks to reduce block index accesses and uses a bitmap to store the distribution of non-zero elements in a block to reduce the storage overhead. A parallel blocking algorithm of BCB-SpTC is designed to divide the binary linear indices into free and contracted indexes to improve the pairing overhead of computational tasks. Then based on the characteristic computation method of TCUs, the proprietary filling method of TCUs is designed to overcome the inefficiency of parallel computation of sparse data on TCUs. Finally, experimental results on the A100 dataset show that BCB-SpTC improves the acceleration ratio by$1.1\times$to$21.3\times$over the existing SpTC GPU method.

In network distance measurement, how to estimate the whole network distance data from partially observed samples has attracted lots of attention because of its significance for network performance evaluation. Matrix completion becomes the most effective approach. However, the two-dimension matrix can only capture the spatial features in the network distance data while ignoring the temporal features. To conquer the problem, few recent studies begin to model the network distance data as a three-dimension tensor and propose tensor completion approaches for distance estimation. Although promising, existing tensor completion approaches still suffer the problem of low recovery accuracy and high measurement cost because they ignore the history priority information. To fully utilize both spatial and temporal features hidden in the distance data, this paper formulates a novel History Priority Enhanced Tensor Completion (HPETC) for distance estimation as a weighted tensor nuclear norm minimization problem where the weight is defined based on the history subspaces information. To solve the weighted tensor nuclear norm minimization problem, we firstly transform it into a factorization-based Frobenius norm minimization problem to avoid costly T-SVD computations, and then propose an iterative algorithm to solve the transformed problem. We further derive a theoretical sampling bound that is lower than the existing sampling bound, thus leads a lower measurement cost. We demonstrate the effectiveness of the proposed algorithm by conducting extensive experiments using two real network distance datasets. The result shows that the proposed algorithm can not only improve the estimation accuracy but also reduce the sampling complexity compared to the state-of-the-art approaches.

Many fields of scientific simulation, such as chemistry and condensed matter physics, are increasingly eschewing dense tensor contraction in favor of sparse tensor contraction. In this work, we center around binary sparse tensor contraction (SpTC) which has the challenges of index matching and accumulation. To address these difficulties, we present GSpTC, an efficient element-wise SpTC framework on CPU-GPU heterogeneous systems. GSpTC first introduces a fine-grained partitioning strategy based on element-wise tensor contraction. By analyzing and selecting appropriate dimension partitioning strategies, we can efficiently utilize the multi-threading parallelism on GPUs and optimize the overall performance of GSpTC. In particular, GSpTC leverages multi-threading parallelism on GPUs for the contraction phase and merging phase, which greatly accelerates the computation phase in sparse tensor contraction computations. Furthermore, GSpTC employs parallel pipeline technology to hide the data transmission time between the host and the device, further enhancing its performance. As a result, GSpTC achieves an average performance improvement of 267% compared to the previous state-of-the-art framework Sparta.

Cyber-physical systems (CPSs) are intricate systems integrating both physical and computational components. When these components fail due to malfunction or cyber-attack, potentially leading to significant damage or even collapse of the network topology. Cyber resilience, defined as the capability of a network to restore its function and structure after component failures, is crucial for ensuring that CPSs can sustain their operational capabilities in the face of complex disturbances. Recently, CPS resilience has garnered increasing attention, leading to the development of various resilience recovery methods. However, most existing studies address network and physical layer resilience in isolation, which hampers the ability to implement adaptive resilience recovery decisions across different systems. To overcome these limitations, we propose a multi-layered cascaded resilient recovery framework grounded in deep reinforcement learning. Initially, we synthesize the complex interactions between the information and physical layers in CPS resilience recovery from a global perspective, modeling the interrelations within CPSs. Subsequently, we introduce a hybrid resilient recovery strategy, encompassing both horizontal and vertical resilient recovery. The correlation matrix is used to partition the system into horizontal and vertical resilience slices. The resilient recovery strategy is subsequently modeled as an optimization problem using these slices. Following this, the Deep Recurrent Q-learning (DRQL) algorithm is introduced to implement the resilient recovery strategy in CPSs. While DRQL exhibits strong adaptability, it may lead to the sparse selection of critical samples, thereby hindering the learning process and convergence on essential experiences. To address this issue, we further develop the RR-DRQL algorithm, designed to identify the optimal CPS resilient recovery strategy. The RR-DRQL algorithm is rigorously proven to converge to the optimal solution through extensive theoretical analysis. Comprehensive experiments demonstrate that the RR-DRQL algorithm surpasses existing resilience recovery methods by 3.8%–25% regarding resilient policy recovery performance across realistic scenarios and various simulation platforms.

The efficiency of production and equipment maintenance costs in the Industrial Internet of Things (IIoT) are directly impacted by equipment lifetime, making it an important concern. Mobile edge computing (MEC) can enhance network performance, extend device lifetime, and effectively reduce carbon emissions by integrating energy harvesting (EH) technology. However, when the two are combined, the coupling effect of energy and the system's communication resource management pose a great challenge to the development of computational offloading strategies. This paper investigates the problem of maximizing the energy efficiency of computation offloading in a two-tier MEC network powered by wireless power transfer (WPT). First, the corresponding mathematical models are developed for local computing, edge server processing, communication, and EH. The proposed fractional problem is transformed into a stochastic optimization problem by Dinkelbach method. In addition, virtual power queues are introduced to eliminate energy coupling effects by maintaining the stability of the battery power queues. Next, the problem is then resolved through the utilization of both Lyapunov optimization and convex optimization method. Consequently, a wireless energy transmission-based algorithm for maximizing energy efficiency is proposed. Finally, energy efficiency, an important parameter of network performance, is used as an indicator. The excellent performance of the EEMA-WET algorithm is verified through extensive extension and comparison experiments.

As the demand for Big Data analysis and artificial intelligence technology continues to surge, a significant amount of research has been conducted on cloud computing services. An effective workflow scheduling strategy stands as the pivotal factor in ensuring the quality of cloud services. Dynamic voltage and frequency scaling (DVFS) is an effective energy-saving technology that is extensively used in the development of workflow scheduling algorithms. However, DVFS reduces the processor's running frequency, which increases the possibility of soft errors in workflow execution, thereby lowering the workflow execution reliability. This study proposes an energy-aware reliability enhancement scheduling (EARES) method with a checkpoint mechanism to improve system reliability while meeting the workflow deadline and the energy consumption constraints. The proposed EARES algorithm consists of three phases, namely, workflow application initialization, deadline partitioning, and energy partitioning and virtual machine selection. Numerous experiments are conducted to assess the performance of the EARES algorithm using three real-world scientific workflows. Experimental results demonstrate that the EARES algorithm remarkably improves reliability in comparison with other state-of-the-art algorithms while meeting the deadline and satisfying the energy consumption requirement.

Convolutional neural networks (CNNs) are widely utilized in intelligent edge computing applications such as computational vision and image processing. However, as the number of layers of the CNN model increases, the number of parameters and computations gets larger, making it increasingly challenging to accelerate in edge computing applications. To effectively adapt to the tradeoff between the speed and accuracy of CNNs inference for smart applications. This paper proposes an FPGA-based adaptive CNNs inference accelerator synergistically utilizing filter pruning, fixed-point parameter quantization, and multi-computing unit parallelism called APPQ-CNN. First, the article devises a hybrid pruning algorithm based on the L1- norm and APoZ to measure the filter impact degree and a configurable parameter quantization fixed-point computing architecture instead of floating-point architecture. Then, design a cascade of the CNN pipelined kernel architecture and configurable multiple computation units. Finally, conduct extensive performance exploration and comparison experiments on various real and synthetic datasets. With negligible accuracy loss, the speed performance of our accelerator APPQ-CNN compares with current state-of-the-art FPGA-based accelerators PipeCNN and OctCNN by 2.15x and 1.91x, respectively. Furthermore, APPQCNN provides settable fixed-point quantization bit-width parameters, filter pruning rate, and multiple computation unit counts to cope with practical application performance requirements in edge computing.

Through the attention mechanism, Transformer-based large-scale deep neural networks (LSDNNs) have demonstrated remarkable achievements in artificial intelligence applications such as natural language processing and computer vision. The matrix-matrix multiplication operation (MMMO) in Transformer makes data movement dominate the inference overhead over computation. A solution for efficient data movement during Transformer inference is to embed arithmetic logic units (ALUs) into the memory array, hence an accelerator-in-memory architecture (AIMA). Existing work along this direction has not considered the heterogeneity of parallelism and resource requirements among Transformer layers. This increases the inference latency and lowers the resource utilization, which is critical for the embedded systems domain. To this end, we propose HAIMA, a hybrid AIMA and the parallel dataflow for Transformer, which exploit the cooperation between SRAM and DRAM to accelerate different MMMOs. Compared to the state-of-the-art Newton and TransPIM, our proposed hardware-software co-design achieves 1.4x-1.5x speedup, and solves the problem of resource under-utilization when DRAM-based AIMA performs the light-weight MMMOs.

Incoherent optical DNN accelerators (OAs) are booming thanks to unparalleled performance-per-watt and excellent scalability. To boost their innovative development, a recent work revolutionarily proposed automatic OA search. However, the robustness and the acceleration performance of the generated OAs are below expectation, because the impacts of inter-tile data transfer and fabrication process & thermal variations (i.e., PTVs) on OAs were ignored. Both hinder OA design automation from being a reality. To resolve theses challenges, we develop FIONA, a novel framework for Fine-grained Incoherent Optical DNN Accelerator search towards both superior acceleration efficiency and inference robustness. Compared against 5 state-of-the-art incoherent OAs on 9 DNN benchmarks, extensive experiments and ablation studies validate the effectiveness of FIONA, achieving up to 198.01× acceleration efficiency improvement based on guaranteed robustness.

Multi-streamed Solid-State Disks (SSDs) have attracted increasing adoption in modern flash storage devices. Despite their excellent promise, effective flash resource allocation is still limiting both their achievable I/O performance and practical implementation. To this end, we develop the first-of-its-kind framework dubbed RLAlloc, which for the first time demonstrates deep Reinforcement Learning-assisted resource Allocation for boosting both I/O throughput and QoS performance of multi-streamed SSDs. Extensive experiments consistently validate the effectiveness of RLAlloc, improving up to 39.9% on I/O throughput and 44.0% on QoS performance over the state-of-the-art competitors.

Learning effective strategies in sparse reward tasks is one of the fundamental challenges in reinforcement learning. This becomes extremely difficult in multi-agent environments, as the concurrent learning of multiple agents induces the non-stationarity problem and sharply increased joint state space. Existing works have attempted to promote multi-agent cooperation through experience sharing. However, learning from a large collection of shared experiences is inefficient as there are only a few high-value states in sparse reward tasks, which may instead lead to the curse of dimensionality in large-scale multi-agent systems. This paper focuses on sparse-reward multi-agent cooperative tasks and proposes an effective experience-sharing method, Multi-Agent Selective Learning (MASL), to boost sample-efficient training by reusing valuable experiences from other agents. MASL adopts a retrogression-based selection method to identify high-value traces of agents from the team rewards, based on which some recall traces are generated and shared among agents to motivate effective exploration. Moreover, MASL selectively considers information from other agents to cope with the non-stationarity issue while enabling efficient training for large-scale agents. Experimental results show that MASL significantly improves sample efficiency compared with state-of-the-art MARL algorithms in cooperative tasks with sparse rewards.

As a novel intelligent sensing paradigm, spatial crowdsourcing has received extensive attention. Task assignment is a key issue in spatial crowdsourcing. In practice, tasks are unevenly distributed in time and space. Accordingly, the problem of cross task assignment attracts growing attention in both industry and academia. Although there has been a research on this problem, it focuses only on maximizing total revenues for inner platforms. Therefore, it can also be improved to bring a multi-win situation for outer workers and task requesters as well as the inner platform. Inspired by this, we first formulate a new cross dynamic task assignment (CDTA) problem by introducing the reputation scores of workers, and prove it to be NP-hard. For the CDTA problem, a hybrid batch-based framework is presented on the basis of a new cross-platform incentive mechanism and a hybrid batch processing strategy, which are efficient in solving the problem of uneven spatial and time distribution of tasks, respectively. After that, a KM-based algorithm and a density-aware greedy algorithm are proposed to gain an accurate assignment result of tasks in each batch and good performance, respectively. Furthermore, the CDTA problem is modeled as a potential game that is proven to have at least a pure Nash Equilibrium theoretically. Last but not least, a game-theoretic approach is developed to maximize the revenues of the inner platform and outer workers at the same time. Extensive experiments on both real and synthetic datasets are conducted to demonstrate the effectiveness and efficiency of the proposed algorithms.

Bipartite graph is an important data structure that widely exists in disease prevention and control, community detection, and other real-life applications. In a bipartite graph, edges not only connect entries of different types but also are bridges of different communities in the above applications. However, research to date has not yet focused on edge importance in bipartite graphs. Inspired by this, we study a new problem of top-k edge search in bipartite graphs with the goal of finding k most important edges for a given bipartite graph; these edges are crucial bridges among communities. In particular, we introduce the measure of ego-betweenness for evaluating the importance of edges. To handle this problem effectively, a lazy bound-based algorithm is first proposed by integrating an upper bound pruning strategy. After that, to further get better efficiency, a greedy bound-based heuristic algorithm is explored on the basis of a tighter upper bound which contributes to reducing redundant computation for calculating ego-betweenness. Last but not least, two parallel techniques with different levels of granularity, called P -src and P -task, are respectively introduced to further improve the search efficiency. The experimental results on both real-world and synthetic graphs demonstrate the efficiency and scalability of the proposed algorithms.

We design a computational framework for ion implantation into silicon that can simulate the reaction process of cascade collisions in silicon and the subsequent annealing process. Since simulating materials requires large-scale particle data and high-precision simulation requirements, which often take up a lot of computing time, the conventional computing process needs to be accelerated in parallel to solve current problems. However, the current field of ion implantation focuses on the development of algorithms, often neglecting the design of parallel algorithms, and has not tested them on large-scale supercomputing clusters. Based on Tianhe new generation high-performance computers, this paper proposes a parallel computing framework for numerical simulation of ion implanted silicon based on multi-core processor architecture. By optimizing the data structure, the local continuity of the data is enhanced to improve the overall communication efficiency. Divide three-dimensional data for processes and adapt to MPI parallel computing strategies. Finally, a multi-level parallel computing strategy based on OpenMP is added to improve computing efficiency. After being deployed on the Tianhe new generation high-performance computer platform, experimental results show that on a single node, the multi-level parallel computing framework can increase the computing speed by 20.81X. After conducting tests at different scales, we can see that the parallel efficiency can be maintained at a high level.

Today’s scientific simulations are confronting seriously limited I/O bandwidth, network bandwidth, and storage capacity because of immense volumes of data generated in high-performance computing systems. Data compression has emerged as one of the most effective approaches to resolve the issue of the exponential increase of scientific data. However, existing state-of-the-art compressors also are confronting the issue of low throughput, especially under the trend of growing disparities between the compute and I/O rates. Among them, embedded coding is widely applied, which contributes to the dominant running time for the corresponding compressors. In this work, we propose a new kind of embedded coding algorithm, and apply it as the backend embedded coding of ZFP, one of the most successful lossy compressors. Our embedded coding algorithm uses bit groups instead of bit planes to store the compressed data, avoiding the time overhead of generating bit planes and group tests of bit planes, which significantly reduces the running time of ZFP. Our embedded coding algorithm can also accelerate the decompression of ZFP, because the costly procedures of the reverse of group tests and reconstructing bit planes are also avoided. Moreover, we provide theoretical proof that the proposed coding algorithm has the same compression ratio as the baseline ZFP. Experiments with four representative real-world scientific simulation datasets show that the compression and decompression throughput of our solution is up to 2.5× (2.1× on average), and up to 2.1× (1.5× on average) as those of ZFP, respectively.

Modern blockchains supporting smart contracts implement a new form of state machine replication with a trusted and decentralized paradigm. However, inefficient smart contract transaction execution severely limits system throughput and hinders the further application of blockchain.

Deep learning on point clouds has attracted increasing attention for various emerging 3D computer vision applications, such as autonomous driving, robotics, and virtual reality. These applications interact with people in real-time on edge devices and thus require low latency and low energy. To accelerate the execution of deep neural networks (DNNs) on point clouds, some customized accelerators have been proposed, which achieved a significantly higher performance with reduced energy consumption than GPUs and existing DNN accelerators.

Deep neural networks have emerged as an important tool for computer-aided diagnosis. However, deep models for skin lesion classification still face the challenges of intra-class variation and inter-class similarity, as well as data imbalance. To address these challenges, in this paper, we propose a balanced supervised contrastive learning (BSCL) approach for the skin lesion classification task. Our model consists of two branches for supervised contrastive learning and classification, respectively. The introduced supervised contrastive learning branch helps the network to learn more discriminative representations. Moreover, we design both a category -averaging strategy which averages the instances of every class in a mini-batch, and a category-complement strategy which makes all categories to appear in each mini-batch, to balance the influence from different skin lesion categories. Besides, we introduce a multi-weighted classification loss to learn a balanced classifier. Extensive experiments on two benchmarks demonstrate that our approach is able to learn strong feature representations and achieve state-of-the-art skin lesion classification performance.

Inspired by the huge success of generative adversarial networks (GANs), GAN-based night-to-day translation methods have achieved excellent results. However, these methods have not been well visualized and understood, and thus cannot find the characteristics about their night-to-day translation performance. To this end, we present a simple clustering-based parsing approach to effectively understand the internal representations of the GAN-based night-to-day translator. In particular, we first cluster the internal representations of specific layers of the translator into a number of classes. Then, according to the proposed selection strategy, the class that has the most significant impact on the translation performance can be identified. Through experiments on three publicly available datasets, we find the answer of the question (title) is that the characteristics at the junction of bright and dark regions affect the performance of night-to-day translation.

Summary With the rapid development of the internet, cyberspace security issues have become increasingly prominent. The importance of constructing a cyberspace security system is self‐evident, but compared with attackers, defenders in cyberspace are in a castle‐like passive defense state in most cases. Therefore, building a reliable, accurate, timely, and active defense system is challenging. The key is to accurately focus on defense priorities, the anticipation of attackers who will likely succeed, and blocking attacks in a timely manner. In this article, we propose an active defense model based on the interaction of situational awareness and firewalls. First, by biasing the integrity, confidentiality, and availability of assets to get the score of assets, and using the Common Vulnerability Scoring System to assess the threat level of assets, we combine the two to determine the maximum system damage that the asset will suffer if it is lost, and then focus on defense. Meanwhile, log analysis of the network situational awareness platform can predict successful attackers, and then the linked firewall strategy can block these attacks in time before the attackers obtain attack gains. After that, we force the attackers to give up their attacks on the target by increasing the attack cost. We compared our model with iptables auto‐blocking and nginx auto‐blocking, and our model excelled them across the board in terms of comprehensiveness and false positive rate. The experimental results verify thar our active defense model proposed in this article can better reduce the defense cost and increase the attack cost, thus achieving the relatively defense goal.

Virtual machine (VM) replication is a critical task in any cloud computing platform to ensure the availability of the cloud service for the end user. In this task, one primary VM resides on a physical machine (PM) and one or more replicas reside on separate PMs. In cloud computing, VM placement (VMP) is a well‐studied problem in terms of different goals, such as power consumption reduction. The VMP problem can be solved by using heuristics, namely, first‐fit and meta‐heuristics such as the genetic algorithm. Despite extensive research into the VMP problem, there are few works that consider VM replication when choosing a VMP. In this context, we proposed studying the problem of optimal VMP considering VM replication requirements. The proposed work frames the problem at hand as a multiobjective problem and adapts a nondominated sorting genetic algorithm (NSGA‐III) to address the problem. VM replicas’ placement should consider several dimensions such as the geographical distance between the PM hosting the primary VM and the other PMs hosting the replicas. In addition, to this end, the proposed model aims to minimize (1) power consumption, (2) performance degradation, and (3) the distance between the PMs hosting the primary VM and its replica(s). The proposed method is thoroughly tested on a variety of computing environments with various heterogeneous VMs and PMs, including compute‐intensive and memory‐intensive environments. The obtained results illustrate the performance disparity between the adapted NSGA‐III and MOEA/D methods and other methods of comparison, including heuristic and meta‐heuristic approaches, with NSGA‐III outperforming other comparison methods. For instance, in memory‐intensive and in heterogeneous environments, the NSGA‐III method’s performance was superior to the first‐fit, next‐fit, best‐fit, PSO, and MOEA/D methods by 58%, 62%, 64%, 55%, and 31%, respectively.

Distributed deep learning systems effectively respond to the increasing demand for large‐scale data processing in recent years. However, the significant investment in building distributed learning systems with powerful computing nodes places a huge financial burden on developers and researchers. It will be good to predict the precise benefit, i.e., how many times of speedup it can get compared with training on single machine (or a few), before actually building such big learning systems. To address this problem, this paper presents a novel performance model on training iteration time for heterogeneous distributed deep learning systems based on the characteristics of the parameter server (PS) system with bulk synchronous parallel (BSP) synchronization style. The accuracy of our performance model is demonstrated by comparing real measurement results on TensorFlow when training different neural networks with various kinds of hardware testbeds: the prediction accuracy is higher than 90% in most cases.

Manual methods to detect and identify potential ground collapse hazards have a high volume of workload and rely on manual skills and experience strongly, with substantial inconsistencies. This paper proposes an automatic recognition method of potential underground hazards and designs an object detection-based algorithm to predict the hidden danger of ground collapse, called GCNet. The GCNet is a ground collapse hazard detection network, which is developed to detect underground hazards, including cavities, poor soil quality, pipelines, and soil background. The GCNet uses the deep residual network ResNet101 and feature pyramid network (FPN) to extract features and a task coordination network (TCN) to identify the category and location of hidden underground hazards. Further, a feature enhancement method based on the regional binary pattern is proposed to improve the accuracy of the proposed model by expanding the ground-penetrating radar (GPR) data by adding multi-level features to GPR images. The experiments are conducted on a large amount of real GPR data and experiment results show that the proposed automatic recognition method can surpass the existing deep learning-based methods in hidden underground hazard recognition and identification.

The blockchain establishes trust by maintaining a distributed appending-only ledger, which is widely applied to the nodes lacking trust in the edge environment. However, the full-replication storage mode of blockchain is a big challenge for resource-constrained edge devices. What is worse, system performance is also affected by the large storage overhead. Existing solutions to reduce blockchain storage overhead often require additional security assumptions, lack incentives, or fail to account for resource heterogeneity. To overcome these limitations, we design an auction-based storage resource allocation scheme. Winners are selected to store blocks, taking into account the block preferences of nodes, and the fairness of the system. Nodes are incentivized by implementing fairness and equity in distributed auctions and data transactions through smart contracts. Finally, extensive experiments show 65%–81% savings in storage overhead compared to fully replicated storage.

Data compression can efficiently reduce the memory and persistence storage cost, which is highly desirable in modern computing systems, such as enterprise, cloud, and High-Performance Computing (HPC) environments. However, the main challenges of existing data compressors are the insufficient compression ratio and low throughput. This paper focuses on improving the compression ratio of state-of-the-art lossy compression algorithms from the view of applications. Besides, we also use the characteristics of the applications to reduce the runtime overhead. To this end, we explore the idea with Adaptive Mesh Refinement (AMR), which is widely adopted as a computational technique to reduce the amount of computation and memory required in scientific simulations. We propose Level Associated Mapping-based Preconditioning (LAMP) to improve the storage efficiency of AMR applications. The main idea is twofold. First, we utilize the high similarities among the adjacent AMR levels to precondition the data prior to compression. Second, AMR has a unique characteristic of grid structures. We utilize grid structures to rebuild a level associated mapping table, which significantly reduces the runtime overhead of LAMP. Thanks to the optimization techniques of General Matrix Multiplication (GEMM), we further accelerate the process of rebuilding AMR hierarchy for LAMP. Besides, we also block multiple adjacent coordinates within a box and further improve cache locality. The experimental results show that the compression ratios of LAMP are improved up to 63.8% compared to directly compressing the data.

Optimal path planning is one of the major bottlenecks for the effective navigation of robots working towards accomplishing complex missions. To overcome the bottleneck and support efficient applications, this paper presentsAAPP, an accelerative and adaptive path planner based on RRT*, a popular path planning algorithm, on GPU, to alleviate four main performance limitations, i.e., bandwidth limitation, load imbalance, high computing complexity, and the choice of parameters. First,AAPPemploys a data storage structure, named simplified compressed sparse rows (SCSR), to compress the large-scale map data and increase the utilization of bandwidth. Second, to exploit the computing performance of GPU, we propose a two-layer parallel framework for RRT* based on SCSR format, named TLRRT*, by using the dynamic parallelism technique. Third, aiming at the problems of parallel load imbalance and high computing complexity in TLRRT*, we further design a two-stage parallel framework, named TSRRT*, that fully exploits hardware heterogeneity (CPU/GPU) by scheduling tasks on CPU and GPU adaptively. Finally, we present optimizations forAAPPto adaptively select execution schemes and parameters. Experimental results on a heterogeneous CPU/GPU machine show thatAAPPyields the speedup up to$22.72\times$over the RRT* algorithm. Compared to the state-of-the-art,AAPPcan handle large-scale datasets and obtain feasible solutions with shorter trajectory lengths.

Tissue-like P systems with channel states are a type of classical membrane systems in which objects transferred among regions are controlled by states placed in the channels between regions. However, an important biological fact is the existence of a “barrier” to the diffusion of signal molecules, which tend to remain confined to some particular micro-habitat. This feature allows quorum sensing to convey information about the physiological state of spatially separated sub-populations. Therefore, in this article, we design a novel class-variant of P systems namedtissue P systems with states in cells(TSIC P systems). Here, each cell contains one and only one state at any moment (the environment has no state), and objects transferred among regions are controlled by states (or a state) that are placed in the corresponding cells (or a cell). We discuss thecomputability theoryof TSIC P systems by showing that Turing universality is acquired by TSIC P systems, which are worked both in a flat maximal parallelism and in a maximal parallelism. In addition, when cell division is considered in TSIC P systems, then tissue P systems with states in cells and cell division (TSICD P systems) are constructed. The (presumed)computational efficiencyof TSICD P systems is reached by offering a uniform solution to the satisfiability problem.

In parallel distributed data processing frameworks like Spark and Flink, task scheduling has a great impact on cluster performance. Though task Scheduling has proven to be an NP-complete problem, a large number of researchers have proposed many heuristic rules to obtain approximate optimal solutions. But most of them ignore the fact that the resource requirements of tasks are dynamically changing during its runtime. Considering the overall task entire lives, the CPU utilization is often lower during the data transfer. Especially for most distributed data processing platforms, data transmission is time-consuming, which usually resulting in low overall CPU utilization. Similarly, network throughput during task calculations is also low in some cases. In this article, we propose a network load variation perception based heuristic task scheduling algorithm, and based on this implement a dual-phase pipeline task scheduler (D2PTS) from the perspective of dynamic resource requirements that aims at maximizing cluster resource utilization, as a supplement to existing data-parallel frameworks. D2PTS divides the states of task into two phases: network-intensive and network-free. To improve the overall resource utilities, this article proposes different algorithms to evaluate the execution time of network sensitive and network free phases respectively. When an executing task is in the network-free phase, D2PTS can additionally schedule a new network-intensive task at the right time. Under this scheduling policy, the two tasks sharing the same CPU core can be executed as a coarse-grained pipeline. This execution method can start tasks earlier and improve resource utilization. Finally, we have implemented our model prototype on Spark 2.4.3 and conducted a number of experiments to evaluate the performance of our model. Experimental results show that D2PTS can not only minimize application makespan, but also improve resource utilization.

This article is concerned with the influence maximization (IM) problem under a network with probabilistically unstable links (PULs) via graph embedding for multiagent systems (MASs). First, two diffusion models, the unstable-link independent cascade (UIC) model and the unstable-link linear threshold (ULT) model, are designed for the IM problem under the network with PULs. Second, the MAS model for the IM problem with PULs is established and a series of interaction rules among agents are built for the MAS model. Third, the similarity of the unstable structure of the nodes is defined and a novel graph embedding method, termed the unstable-similarity2vec (US2vec) approach, is proposed to tackle the IM problem under the network with PULs. According to the embedding results of the US2vec approach, the seed set is figured out by the developed algorithm. Finally, extensive experiments are conducted to: 1) verify the validity of the proposed model and the developed algorithms and 2) illustrate the optimal solution for IM under different scenarios with PULs.

Multiobjective-based constraint-handling techniques are popular in evolutionary constrained single-objective optimization. However, most of these techniques run into troubles when dealing with constrained multiobjective optimization problems (CMOPs). That is, they have difficulty optimizing too many objective functions, are ineffective in maintaining population diversity, or are challenged in establishing appropriate additional objective functions. As a remedy to these limitations, we propose a novel technique called NRC for handling CMOPs. The novelty of NRC lies in its three sorting procedures: 1) nondominated sorting; 2) reversed nondominated sorting; and 3) constrained crowding distance sorting, which are performed in sequence to provide driving forces toward the Pareto front (PF) of a transformed unconstrained multiobjective optimization problem (treating the overall constraint violation as an additional objective function), the boundary front, and the constrained PF, respectively. With the combination of these three different forces, NRC can conveniently approach the desired PF from diverse search directions. The effectiveness of NRC is experimentally verified. Also, we incorporate NRC into a two-archive mechanism and develop a novel constrained multiobjective evolutionary algorithm, called NRC2. Comprehensive experiments on 49 benchmark CMOPs and 21 real-world ones demonstrate that NRC2 is significantly superior or comparable to six state-of-the-art constrained evolutionary multiobjective optimizers on most test instances.

Transformer model architectures have recently received great interest in natural language, machine translation, and computer vision, where attention mechanisms are their building blocks. However, the attention mechanism is expensive because of its intensive matrix computations and complicated data flow. The existing hardware architecture has some disadvantages for the computing structure of attention, such as inflexibility and low efficiency. Most of the existing papers accelerate attention by reducing the amount of computation through various pruning algorithms, which will affect the results to a certain extent with different sparsity. This paper proposes the hardware accelerator for the multi-head attention (MHA) on field-programmable gate arrays (FPGAs) with reconfigurable architecture, efficient systolic array, and hardware-friendly radix-2 softmax. We propose a novel method called Four inputs Processing Element (FPE) to double the computation rate of the data-aware systolic array (SA) and make it efficient and load balance. Especially, the computation framework is well designed to ensure the utilization of SA efficiently. Our design is evaluated on a Xilinx Alveo U250 card, and the proposed architecture achieves 51.3×, 17.3× improvement in latency, and 54.4×, 17.9× energy savings compared to CPU and GPU.

With the massive amounts of data generated by industrial Internet of Things (IIoT) devices at all moments, federated learning (FL) enables these distributed distrusted devices to collaborate to build machine learning model while maintaining data privacy. However, malicious participants still launch malicious attacks against the security vulnerabilities during model aggregation. This article is the first to propose Sybil-based collusion attacks (SCA) in the IIoT-FL system for the vulnerabilities mentioned above. The malicious participants use label flipping attacks to complete local poisoning training. Meanwhile, they can virtualize multiple Sybil nodes to make the local poisoning models aggregated with the greatest possibility during aggregation. They focus on making the joint model misclassify the selected attack class samples during the testing phase, while other nonattack classes kept the main task accuracy similar to the nonpoisoned state. Exhaustive experimental analysis demonstrates that our SCA has a superior performance on multiple aspects than the state-of-the-art.

Nowadays, more data are represented and stored by a graph structure, and subgraph matching is a fundamental problem in a variety of scientific machine learning and industrial applications, such as remote sensing image registration, industrial inspection, etc. Due to the nondeterministic polynominal hard (NP-hard) problem of subgraph matching, the explosive growth of graph data, the disadvantages of high energy consumption, and the high overhead of CPU and graphics processing unit (GPU) platforms, computing subgraph matching is becoming more and more challenging. To alleviate this problem, we propose a phased hybrid algorithm to accelerate the enumeration task of subgraph matching, calledPH-CF, based on the CPU-field programmable gate arrays (FPGA) heterogeneous platform. This approach can make full use of the pipeline and data flow mechanism, low power consumption, and configurable characteristics of FPGA. First, the matching order of query vertices automatically selects GraphQL (GQL) or RI methods according to the sparsity of the data (query) graph. Second, a candidate vertex auxiliary data structure set partitioning method is designed to effectively realize the load balance of multiple computing units at the FPGA and CPU host sides. Third, FPGA's pipeline and data flow mechanism is used to accelerate the enumeration phase of subgraph matching. Experimental results on real-world and synthetic datasets show that the performance of thePH-CFoutperforms the state-of-the-arts.PH-CFcan obtain the average performance improvement of up to$16.07\times$,$38.61\times$, and$11.46\times$over CFL, CECI, and DP-iso, respectively. Moreover, our approach has good stability and robustness on various datasets.

The ultrasound standard plane plays an important role in prenatal fetal growth parameter measurement and disease diagnosis in prenatal screening. However, obtaining standard planes in a fetal ultrasound video is not only laborious and time-consuming but also depends on the clinical experience of sonographers to a certain extent. To improve the acquisition efficiency and accuracy of the ultrasound standard plane, we propose a novel detection framework that utilizes both the coarse-to-fine detection strategy and multi-task learning mechanism for feature-fused images. First, traditional manually-designed features and deep learning-based features are fused to obtain low-level shared features, which can enhance the model's feature expression ability. Inspired by the process of human recognition, ultrasound standard plane detection is divided into a coarse process of plane type classification and a fine process of standard-or-not detection, which is implemented via an end-to-end multi-task learning network. The region-of-interest area is also recognised in our detection framework to suppress the influence of a variable maternal background. Extensive experiments are conducted on three ultrasound planes of the first-class fetal examination, i.e., the femur, thalamus, and abdomen ultrasound images. The experiment results show that our method outperforms competing methods in terms of accuracy, which demonstrates the efficacy of the proposed method and can reduce the workload of sonographers in prenatal screening.

Unsupervised bipartite graph learning has been a hotpot in multi-view clustering, to tackle the restricted scalability issue of traditional full graph clustering in large-scale applications. However, the existing bipartite graph clustering paradigm pays little attention to the adverse impact of noisy features on learning process. To further facilitate this part of research, apart from simply reweighting features to depress the noisy ones, we take the first step towards analyzing the induced adverse impact via theoretical and experimental investigations. One crucial finding in this paper is that the existence of noisy features will incur “anchor shift” phenomenon, which deviates the potential representations of anchors and then degrades performance. To this end, we propose a coupled noisy feature filter mechanism with automatically finding feature importance to remedy the anchor shift issue in this paper. Apart from leveraging features, we theoretically analyze the bounds of proposed feature-adaptive bipartite graph's fuzzy membership. Specifically, distinguishing features' discrimination will increase the fuzzy membership to achieve soft partitions against the potential inaccurate absolute relationship. With the afore-mentioned merits, our proposed multi-view bipartite graph clustering with coupled noisy feature filter model (MVBGC-NFF) provides novel and interesting insights on the feature level of anchor shift. The effectiveness and efficiency of MVBGC-NFF are demonstrated on synthetic and real-world datasets with improving clustering performance, increasing fuzzy membership, and filtering noisy features. The code is available onhttps://github.com/liliangnudt/MVBGC-NFF.

Influential community search aims to find cohesive subgraphs (communities) with considerable influence. It is a fundamental graph management operator that can play a crucial role in biological network analysis, activity organization, and other real-life applications. Existing research on influential community search is mainly focused on deterministic graphs with the assumption that influences between entities are certain. This assumption is invalid in many cases because it ignores the uncertainty which is an inherent property of influence. Against this backdrop, in this paper, we introduce an uncertain influential community model, namely$(k, \eta)$-influential community, based on which the influential community search problem over uncertain graphs is formulated. Furthermore, we propose an online approach by integrating a peeling-pruning strategy that can progressively refine the given uncertain graph to find the$(k, \eta)$-influential communities. To further improve the search performance, two novel indexes, ICU-Index and FICU-Index, are developed to organize the$(k, \eta)$-influential communities at different probabilistic intervals. The indexes decompose the probabilistic interval into multiple subintervals and based on this, the$(k, \eta)$-influential communities are divided into different groups in turn. Compared with ICU-Index, FICU-Index requires considerably less space with the introduction of two optimization strategies. These indexes help obtain results of an influential community search problem more efficiently. Extensive experiments on large real and synthetic datasets demonstrate the efficiency and effectiveness of our proposed algorithms.

How to efficiently coordinate multiple readers to work together is critical for high throughput in RFID systems. Existing researchs focus on designing efficient reader scheduling strategies that arrange adjacent readers to work in different time to avoid signal collisions. However, the impact of unbalanced tag number of readers on tag read throughput is still challenging. In RFID systems, the distribution of tags is usually variable and uneven, making the number of tags covered by each reader (i.e., the load) imbalanced. This imbalance leads to different execution time for readers: the heavily loaded readers take longer time to collect all tags, while the other readers whose finish execution earlier have to wait in vain. To avoid this useless waiting and improve the system throughput, this paper focuses on the load balancing problem of multiple readers, which is an NP-hard problem. In this paper, we design heuristic algorithms to adjust readers interrogation regions and efficiently balance their loads. The amazing advantage of our algorithm is that it can be adopted by almost all existing protocols in multi-reader systems, including the reader scheduling protocol, to improve system throughput. Extensive experiments demonstrate that our algorithm can significantly improve the throughput in various scenarios.

RFID systems have been widely deployed for various applications such as inventory control, retail management, and object tracking. In many scenarios, coarse-grained system management of tag categories is more common than traditional system management of individual tags. Existing works on tag category management require knowing category IDs in advance, however, few protocols have been specially designed to obtain category IDs in multi-category RFID systems. In this paper, we proposeCIP, a time-efficient protocol specially designed to collect category IDs in multi-category RFID systems. The challenge in solving this problem is to avoid repeatedly collecting category IDs from multiple tags belonging to the same category. We address this challenge by squeezing tags in the same category into the same time slot and letting them reply to the reader simultaneously. Considering that the reader may receive signals from different tags in a slot, we exploit the Manchester coding to decode the aggregated signals. With this design, CIP collects each category ID only once and thus achieves high time efficiency by avoiding redundant identification of the same category ID. We further enhance CIP by utilizing the lightweight vectors to specify the order of tags’ replying and avoid useless slots. Rigorous theoretical analysis is presented to optimize the performance of CIP and guidelines on how to set optimal parameters to minimize the overall execution time are given. We conduct extensive experiments to evaluate the performance of our protocols. The experimental results show that our protocols can significantly improve time efficiency by 67 percent when compared with the state-of-the-art solutions.

Logs are pervasive in modern computing systems, and are valuable to service and system management. Nevertheless, with the rapidly growing size and complexity of computing systems, the log volume is exploding, which makes automatic log analysis imperative. Generally, in automatic log analysis, the first and fundamental step is log parsing, to which a lot of effort has been devoted. However, in most existing log parsing methods, log messages are merely treated as plain text. In natural language processing (NLP) area, it is a common practice to represent words and sentences with vectors, then the similarity between two words or sentences can be measured by the distance between their vectors. Inspired by these, we put forward a novel log parsing framework, named LPV (LogParser based onVectorization), which performs log parsing by converting log messages and log templates into vectors, with the help of a vectorization method in NLP. LPV incorporates offline and online log parsing. In the offline log parsing, the central idea is to first represent log messages with vectors, so that the similarity between two log messages can be measured by the distance between their vectors, then we cluster log messages via clustering the vectors, and finally we extract log templates from the resultant clusters. By the end of the offline log parsing, each log template is assigned with an average vector, so that in the online log parsing, the similarity between an incoming log message and each log template can also be measured by the distance between their vectors. Extensive experiments have been conducted based on several public log datasets to evaluate LPV with three different vectorization methods. The results demonstrate that, with a proper vectorization method, LPV performs competitive with state-of-the-art log parsing methods, in both effectiveness and efficiency.

The Point Of Interest (POI) sequence recommendation is the key task in itinerary and travel route planning. Existing works usually consider the temporal and spatial factors in travel planning. However, the external environment, such as the weather, is usually overlooked. In fact, the weather is an important factor because it can affect a user’s check-in behaviors. Furthermore, most of the existing research is based on a static environment for POI sequence recommendation. While the external environment (e.g., the weather) may change during travel, it is difficult for existing works to adjust the POI sequence in time. What’s more, people usually prefer the attractive routes when traveling. To address these issues, we first conduct comprehensive data analysis on two real-world check-in datasets to study the effects of weather and time, as well as the features of the POI sequence. Based on this, we propose a model of Dynamic Personalized POI Sequence Recommendation with fine-grained contexts ( DPSR for short). It extracts user interest and POI popularity with fine-grained contexts and captures the attractiveness of the POI sequence. Next, we apply the Monte Carlo Tree Search model (MCTS for short) to simulate the process of recommending POI sequence in the dynamic environment, i.e., the weather and time change after visiting a POI. What’s more, we consider different speeds to reflect the fact that people may take different transportation to transfer between POIs. To validate the efficacy of DPSR , we conduct extensive experiments. The results show that our model can improve the accuracy of the recommendation significantly. Furthermore, it can better meet user preferences and enhance experiences.

Network trouble shooting, failure location, and anomaly detection rely heavily on network traffic measurement data. Due to the lack of measurement infrastructure, the high measurement cost, and the unavoidable transmission loss, network monitoring systems suffer from the problem that the network traffic data are incomplete. This article models the traffic data as a tensor to exploit its strong ability of feature extraction to recover the missing data. Different from traditional tensor completion which relies on tensor factorization, we design a novel Deep Adversarial Tensor Completion (DATC) scheme based on Deep Learning (DL) techniques. DATC is the first scheme that exploits the data reconstruction ability of autoencoder and the power of adversarial training from Generative Adversarial Networks to infer the missing data. Despite that DL techniques achieve great success in the image field, designing an algorithm based on DL techniques to recover the traffic data with missing entries faces additional challenges due to the skewed distribution and the sparsity of traffic data. To conquer these challenges, we propose the use of two techniques, adversarial training and missing data aware convolution. These techniques help DATC to learn the complex features of the traffic data and infer the missing data following the data distribution of traffic data. Our extensive experimental results using two public real-world network traffic datasets and running both offline and online demonstrate that DATC can achieve significantly better recovery accuracy while capturing the data distribution of the traffic data even when the sampling ratio is very low.

Graph convolutional network (GCN) has been extensively applied to recommender systems (RS) and achieved significant performance improvements through iteratively aggregating high-order neighbors to model the relevance between users and items as well as their characteristics. In the aggregation process, GCN models usually give neighbors the same or trainable weights based on implicit features, ignoring explicit ones. In this work, we take the features with explicit meanings or extracted with specific purpose as explicit ones (e.g., temporal features) and the others contained in user-item network as implicit ones (e.g., user preferences). However, some explicit features and knowledge play an essential role in improving the model representation ability and explainability in recommendation systems. To deal with the limitation, we propose a GCN based framework to embed the explicit features or those extracted with explicit intentions in this work. We also provide specific implementations based on two commonly researched features, temporal evolution and popularity bias. Specifically, we first experimentally analyze the popularity bias of the representation learning in RS based on two commonly used GCN models. Secondly, we propose a general framework to weigh neighbors based on explicit features or intentions. Thirdly, we implement a Temporal and Popularity weighted Aggregator (TPA) for GCN. The Interest-Forgetting Curve is utilized to capture temporal evolution as temporal weights and the data-driven Beta distribution is employed to tune the weights based on the node popularity flexibly. At last, we conduct extensive experiments on three real-world datasets to demonstrate the effectiveness of TPA in improving recommendation accuracy and alleviating the popularity bias.

Both accurate and fast mobile recommendation systems based on click behaviors analysis are crucial in e-business. Deep learning has achieved state-of-the-art accuracy and the traditional wisdom often hosts these computation-intensive models in powerful cloud centers. However, the cloud-only approaches put significant computational pressure on cloud servers and increase the latency in heavy-load scenarios. Moreover, existing work often adopts RNN structures to model behaviors that suffer from low processing speed for under-utilization of parallel devices such as GPUs. In this work, we propose an efficient internet behavior-based recommendation framework with edge-cloud collaboration on deep CNNs (CoRec) to improve both the accuracy and speed for mobile recommendation. A novel convolutional interest network (CIN) improves the accuracy by modeling the long- and short-term interests and accelerates the prediction through parallel-friendly convolutions. To further improve the serving throughput and latency, a novel device-cloud collaboration strategy reduces workloads by pre-computing and caching long-term interests in the cloud offline and real-time computation of short-term interests in devices. Extensive experiments on real-world datasets show that CoRec significantly outperforms the state-of-the-art methods in accuracy and has achieved at least an order of magnitude improvement in latency and throughput compared to cloud-only RNN-based approaches for long behaviors.

Detecting anomalous traffic is a crucial task for network management. Although many anomaly detection algorithms have been proposed recently, constrained by their matrix-based traffic data model, existing algorithms often suffer from low detection accuracy. To fully utilize the multi-dimensional information hidden in the traffic data, this paper uses the tensor model for more accurate Internet anomaly detection. Only considering the low-rank linearity features hidden in the data, current tensor factorization techniques would result in low anomaly detection accuracy. We propose a novel Graph-based Tensor Recovery model (Graph-TR) to well explore both low-rank linearity features as well as the non-linear proximity information hidden in the traffic data for better anomaly detection. We encode the non-linear proximity information of the traffic data by constructing nearest neighbor graphs and incorporate this information into the tensor factorization using the graph Laplacian. Moreover, to facilitate the quick building of neighbor graph, we propose a nearest neighbor searching algorithm with the simple locality-sensitive hashing (LSH). Besides only detecting random anomalies, our algorithm can also effectively detect structured anomalies that appear as bursts. We have conducted extensive experiments using Internet traffic trace data Abilene and GÈANT. Compared with the state of art algorithms on matrix-based anomaly detection and tensor recovery approach, our Graph-TR can achieve higher Accuracy and Recall.

Network measurements provide critical inputs for a wide range of network management. Existing network-wide monitoring methods face the challenge of incurring a high measurement cost. Some recent studies show that network-wide measurement data such as end-to-end latency and flow traffic, have hidden spatio-temporal correlations and thus low-rank features. Taking advantage of the low-rank feature, enlightened by tensor model's strong capability of information representation and extracting, this paper studies a novel sparse measurement scheduling problem which selects a proportion of Origin and Destination (OD) pairs to take measurements in the future time slots, while ensuring the data of the remaining un-measured OD pairs be accurately inferred through tensor completion. It is challenging to find the optimal sampling points (OD pairs) without knowing the structure of the future data and also infer the un-measured data in the presence of noise in the measurement samples. To conquer the challenges, we propose several techniques: a tripartite graph to illustrate the relationship between sample locations and tensor factorization, a graph-based sample selection algorithm, and a graph-based robust tensor completion algorithm. We have conducted extensive experiments based on two real network latency monitoring traces (PlanetLab and Harvard) and two other network monitoring traces (including a traffic trace Abilene and a throughput trace WS-Dream). Our results demonstrate that, even with a sampling ratio of less than 5%, our scheme can accurately obtain the complete network-wide monitoring data by inferring the missing ones based on the samples taken. To achieve similar recovery performance, the best peer tensor completion algorithm needs a significantly larger number of samples, with the sampling ratio up to 25-150 times ours.

Route travel time estimation (RTTE) is crucial in intelligent transportation systems. Performing aggregation is a fundamental operation in RTTE and is widely used in the traffic prediction and route calculation stages. Observations have revealed that aggregation operations in RTTE are influenced by the road network structure and aggregation requests, resulting in irregular data access, redundant processing, and workload imbalance. Existing architectures have not addressed these issues effectively. In this study, we begin by characterizing the execution pattern of performing aggregation operations on an Intel Core CPU. Guided by these characterizations, we propose an aggregation accelerator that utilizes a hierarchical reduction architecture (HRA) to perform aggregations in RTTE efficiently. Specifically, we construct an inverted table based on the road network and aggregation requests. Building upon the concept of hierarchical reduction, we design an HRA to accelerate aggregation operations which reduces irregular data access and eliminates redundant processing. Additionally, we introduce a reconfiguration mode for HRA to mitigate workload imbalance issues. Compared to a benchmark method executed on an Intel Core CPU, our design achieves an average$10\times$speedup.

Network traffic anomaly detection is critical for advanced network applications. However, network traffic monitoring data arrive in a streaming fashion and could be infinite, which makes the offline algorithms that attempt to store the entire stream monitoring data for analysis not scalable. To well utilize the strong ability of tensor model, we use a tensor to represent the prior non-anomalous traffic matrices and propose a novel unsupervised anomaly detection framework that can be used to detect anomalies in a streaming fashion by making only one pass over the data while utilizing limited storage. In the framework, we propose a succinct tensor sketch to maintain, in a streaming model, the subspace that can well represent all prior non-anomalous data detected. Using the subspace, anomalies in each new incoming traffic monitoring data can be quickly detected based on a simple outlier score calculation. Further, we prove that the tensor sketch is mergeable. Exploiting this property, we propose a distributed anomaly detection framework in which the distributed node only needs to upload its succinct tensor sketch instead of the raw monitoring data to the central node to calculate the global subspace of the whole network, which greatly saves the transmission cost. We theoretically prove that our tensor sketch based anomaly detection algorithm compares favorably with the offline approach which calculates the subspace based on expensive global Singular Value Decomposition (SVD). The experimental results demonstrate the effectiveness and efficiency of our approach over other popular online anomaly detection algorithms.

In this paper, we present COFFEE, cross-layer optimization for fast and efficient executions of the Sinkhorn-Knopp (SK) algorithm on HPC systems with clusters of compute nodes by exploring some architectural features of the system. By analyzing the performance of a typical implementation of the SK algorithm on such a system, a huge performance gap is observed between the row rescaling and column rescaling of the algorithm, where the latter requires much more time than the former. We also found that the costly MPI communication of the column rescaling seriously hinders the exploitation of parallelism. By observing and leveraging unique architectural characteristics across different system optimizations, such as column rescaling redesign, data blocking, micro-kernel design, enhanced intra-node and inter-node communication in MPI, etc., COFFEE is able to explore cross-layer optimization opportunities that enable fast and efficient execution of the SK algorithm. Our experimental results show that COFFEE provides up to 7.5X with an average of 2.0X performance improvement over the typical implementation on a single node, and up to 2.9X with an average of 1.6X performance improvement over the state-of-the-art MPI Allreduce algorithms on Tianhe-1 supercomputer.

Analysis of multi-dimensional data, especially tensor decomposition, which extracts latent information, is becoming considerably popular. Although multi-dimensional sparse data is typically processed on multi-core processors, developing highly optimized GPU-basedSparseTensorMatrixChainMultiplication (SpTMCM) is challenging. The purpose of this paper is to investigate a novel approach named SpTMCM and to explore the discovery of SpTMCM coupled with the emerging computing core, Tensor Core Unit (TCU). In contrast to prior work, the proposed novel approach enables a uniform storage format and optimization approach for SpTMCM. We design a hybrid tensor format based on multi-dimensional tiling that divides the tensor depending on the tile threshold to address the inefficient memory accesses caused by the irregular nonzero distribution of the sparse tensor. Further, we develop a TCU-based tensor parallel algorithm with our novel approach to increase the memory bandwidth. Compared to state-of-the-art works, our method achieves$1.16\sim 24.12\times$speedup for SpMTTKRP and$5.07\sim 7.15\times$speedup for SpTTMChain across NVIDIA A100 GPU on a range of real-world sparse tensors.

Tip decomposition has a pivotal role in mining cohesive subgraphs in bipartite graphs by computing the tip number of each vertex in accordance with the non-trivial motif butterfly ((2,2)-biclique). It has been a popular research topic with applications in document clustering, spam group detection, and analysis of affiliation networks. In such applications, the graphs are not only large, but they evolve quickly with new edges being continuously added/deleted. While existing centralized techniques could solve the tip decomposition problem for static bipartite graphs, they are not efficient for maintaining the tip numbers of vertices on large-scale graphs. In this paper, we study butterfly analysis problems on bipartite graphs in a distributed environment with the vertex-centric model. We first extend a centralized butterfly counting algorithm to a distributed version, called DBCA. An ingenious message aggregation strategy is designed to reduce massive redundant messaging and avoid the memory overflow problem while processing large-scale graphs. Based on the results of DBCA, we develop a distributed tip decomposition algorithm (DTDA) to get the tip number of each vertex in parallel. Finally, to maintain the tip numbers of vertices efficiently while graphs evolve over time, we explore a distributed tip maintenance algorithm (DTMA) along with a novel task-split strategy. Specifically, for an updated edge (insertion/deletion), several sub-tasks will be generated in line with the topology structure of the original bipartite graph. To our best knowledge, this is the first study to process the butterfly analysis problems in a distributed environment. In addition, comprehensive experiments have been conducted on real-world bipartite graphs. The experiment results demonstrate that our proposed algorithms are efficient and scalable.

The service resource allocation strategy optimization problem has always been a hot issue in mobile edge computing (MEC). In this paper, the problem is formulated as a long-term quality of service (QoS) improvement problem while satisfying the budget of MEC service provider (MSP). Since it is very unrealistic to accurately obtain the request information of user equipments (UEs) over a long time, we first transform the original problem into a series of real-time linear programing sub-problems by using Lyapunov optimization method, and propose a centralized algorithm to determine the resource allocation strategies. However, since the sub-problems are still NP-hard problems, it is a huge challenge to determine the strategies for all UEs with the centralized algorithm in a large scale MEC environment. Thus, we then formulate the sub-problems as an N players non-cooperative game, prove that there exists a Nash equilibrium, and develop two iterative algorithms to find the Nash equilibrium while determining the strategies. Experimental results show that the algorithms can take into account QoS and budget of MSP at the same time, and perform better compared to five other common schemes.

By offloading computation-intensive tasks to the edge cloud, mobile edge computing (MEC) has been regarded as an effective technology for enhancing computational capacity and extending the battery lifetime of mobile devices (MDs). However, due to the limitation of bandwidth and computing resources in MEC, unreasonable task offloading might lead to intensive resource competition, which recedes the performance gains benefit from offloading. When the tasks are latency-sensitive, a proper task offloading strategy is more important. Considering the heterogeneous delay constraints and resource competition comprehensively, we aim at minimizing the energy consumption of MDs subject to the individual delay constraints of tasks by jointly optimizing the task offloading and resource allocation in terms of wireless channel and remote computation capacity in a multi-MD MEC system in this paper. Due to the complexity of the primal optimization problem, a heuristic algorithm is devised. In the algorithm, a subset of tasks to be offloaded is incrementally constructed, and the corresponding offloading sub-problem is then repeatedly solved for this task subset using a two-stage algorithm until the total energy consumption can no longer be further reduced. The first stage of solving the sub-problem is to find the optimal full offloading scheme for the to-offload tasks, which is proved to be a convex optimization problem. For the task subset without a full offloading solution, an effective iterative algorithm is employed in the second stage where the channel allocation and computing resource allocation are optimized alternately. A great number of experiments are given to verify the performance of the proposed algorithm. We observe that the heuristic algorithm shows different performance when adopting different task ordering schemes. The proposed heuristic algorithm is evaluated against three reference schemes, and the results show that it can save up to 14.20% of energy consumption while guaranteeing the delay requirements of all tasks.

Detecting anomalies in logs is crucial for service and system management, since logs are widely used to record the runtime status, and are often the only data available for postmortem analysis. Since anomalies are usually rare in real-world services and systems, a common and feasible practice is to mine or learn normal patterns from logs, and deem those violating the normal patterns as anomalies. As log sequences are a kind of time series data, RNN (Recurrent Neural Network) and its variants have been extensively employed to capture the normal patterns. Nevertheless, the sequential nature of RNN and its variants makes them hard to parallelize and capture long-term dependencies, which may hinder their performance. To address this issue, in this paper we propose Loader, a novel semi-supervisedloganomalydetector based on Transformer, because the Transformer architecture eschews recurrence and is able to draw global dependencies. Loader leverages the Transformer encoder to capture normal patterns from normal log sequences. When detecting, it gives a set of candidate log templates, that may appear after the input log substring under normal conditions. If the template of the actual next log message is not within the candidate set, this implies an anomaly. Previous similar methods select the most possible$k$log templates as candidates in any case, so the performance is sensitive to$k$, and it is nontrivial to pick a proper$k$. To alleviate this, we design a more flexible and robust ‘top-$p$’ algorithm, which determines the candidate set based on the cumulative probability of the most possible log templates. Extensive experiments are conducted based on three public log datasets, the experimental results validate the effectiveness and competitiveness of our approach.

Graph partitioning is one of the fundamental problems in many graph-based applications and systems. It enables the division of a graph into smaller sub-graphs for subsequent parallel processing, reducing the processing latency of the graph. The critical path of a graph is the logical path with the longest delay from input to output. The processing time of the graph mainly depends on the delay incurred by the critical path, independent of other paths with small delays. Therefore, it can reduce the processing time of the graph by protecting the critical path of the graph from partition. However, existing approaches to graph partitioning only focus on metrics such as minimum cut and partition balance. As a result, the critical paths of graphs may be destroyed in the partitioning procedure. To address this problem, we present a critical path awareness approach, namely path-metis, to protect the critical paths and alleviate the processing latency after graph partitioning. In path-metis, two efficient strategies, including Slack and critical path fix strategies, are introduced. The Slack strategy, which incorporates critical path information into the weights of DAG, is used as pre-processing before traditional multi-level partitioning methods, like Metis. Then, for the generated partitioning scheme, the critical path fix strategy is proposed to further protect critical paths from being cut. We demonstrate the effectiveness of our approach on both real and synthetic datasets. From the experimental results, compared to Metis, our method improves critical path performance by 17.70%.

The legal charge prediction task aims to judge appropriate charges according to the given fact description in cases. Most existing methods formulate it as a multi-class text classification problem and have achieved tremendous progress. However, the performance on low-frequency charges is still unsatisfactory. Previous studies indicate leveraging the charge label information can facilitate this task, but the approaches to utilizing the label information are not fully explored. In this paper, inspired by the vision-language information fusion techniques in the multi-modal field, we propose a novel model (denoted as LeapBank) by fusing the representations of text and labels to enhance the legal charge prediction task. Specifically, we devise a representation fusion block based on the bilinear attention network to interact the labels and text tokens seamlessly. Extensive experiments are conducted on three real-world datasets to compare our proposed method with state-of-the-art models. Experimental results show that LeapBank obtains up to 8.5% Macro-F1 improvements on the low-frequency charges, demonstrating our model's superiority and competitiveness.

Graph neural networks (GNNs), which extend conventional deep learning technologies to process graph-structured data, have shown its powerful graph representation learning ability. Existing typical GNNs utilize neighborhood message passing mechanism based on neural networks that updates target vertex representations by aggregating feature messages from neighboring source vertices. To accelerate the computations of GNNs, some customized accelerators, which follow the neighborhood aggregation computation pattern for each vertex, have been proposed. Through analysis, we observe that a naive implementation of the neighborhood aggregation results in redundant computations and communications.In this paper, we propose a novel redundancy-eliminated GNN accelerator, shortly termed as ReGNN. ReGNN is supported by an algorithm and architecture co-design. We first propose a dynamic redundancy-eliminated neighborhood message passing algorithm for GNNs. Then a novel architecture is designed to support the proposed algorithm and transform the redundancy elimination into performance improvement. ReGNN is also a configurable pipelined architecture that can be configured to support different GNN variants. In terms of the same computations, ReGNN provides the same accuracy as traditional GNNs. To the best of our knowledge, ReGNN is the first accelerator that can eliminate computation redundancy in GNNs. Our proposed ReGNN system gains an average of 9.1× speedup and 8.9× energy efficiency over state-of-the-art GNN accelerators.

Canonical Polyadic Decomposition (CPD) of sparse tensors is an effective tool in various machine learning and data analytics applications, in which sparse Matricized Tensor Times Khatri-Rao Product (MTTKRP) is the major performance bottleneck. To overcome this bottleneck and support efficient applications, this paper presents HPSpTM, an efficient sparse MTTKRP framework, to exploit the multi-level parallelism and reusability on heterogeneous HPC systems. HPSpTM incorporates: (1) a multi-level matrix-driven tiling engine that leverages the process- and thread-level parallelism of the underlying platform and data reusability based on the derived factor matrix-driven MTTKRP algorithm; (2) a tensor-driven parallel execution that enables buffering-aware scheduling and pipeline scheduling to optimize the performance in the tile granularity; (3) a partition-aware light weight data storage that exploits better data locality based on the proposed hierarchical and fine-grained execution; and (4) a performance auto-tuning technique that offers large flexibility for tile size auto-adjusting across various input datasets based on a designed runtime model. Our experiments show that HPSpTM on a Nvidia Tesla P100 obtains the average performance improvement of up to 76.46% over the state-of-the-arts, and HPSpTM achieves the speedup of up to 15.39× when scaling from 8 to 128 core groups, corresponding to processes, on the Sunway TaihuLight supercomputer.

As a typical most-to-most connected quasi-biclique model, k-biplex is a superset of bicliques, which allows nodes on each side of a fully connected subgraph to lose at most$k$connections. In this paper, we investigate the maximum biplex search problem for the first time. The goal here is to find a k-biplex with the maximum number of edges and we have proved that the problem is NP-hard. It is widely used in fraudulent reviewer group detection, gene expression analysis, social recommendation, and other real-life applications. To solve this problem, a maximum k-biplex search algorithm (MBS) is first presented by integrating two pruning strategies, including degree-based and 2-hop-based pruning. In addition, we define a new dense subgraph over bipartite graphs,$\langle x, y\rangle$-core, and develop a core-based maximum k-biplex search algorithm (MBS-Core) which can significantly reduce the search space with the introduction of a core-based graph reduction technique. In particular, it only needs to search these cores instead of the entire graph to obtain the maximum k-biplex. Moreover, a parallel algorithm and a heuristic algorithm are developed to achieve better query performance on larger-scale bipartite graphs. Extensive experiments have been conducted on real-life and synthetic datasets to verify the efficiency and effectiveness of the proposed algorithms. Our results show that MBS-Core is up to 3 orders of magnitude faster than the existing approaches.

Set label-constrained reachability (SLCR) query in edge-labeled graphs is a building block of many graph-based applications. Formally, given two sets$S$and$T$of source and target vertices and a label set (, it returns all reachable vertex pairs (s, t) under the constraint of (, where$s$∊$S$and$t$∊T. There have been abundant index-based approaches to be applied to process the SLCR query. However, distributed approaches are desirable to process large-scale graphs because of the advantages of good scalability and real-time response. Now, there is no efficient distributed approach to the SLCR query. Most index-based approaches face limitations in terms of index construction and query performance when being extended to the distributed environment for processing large-scale graphs. To alleviate these problems, we first build a boundary graph-based index (BoundG) to reduce the time overhead of index construction. Consider the query performance of the BoundG-based approach has no noticeable improvement. We further construct a novel two layers 2-hop index (TL2hop), and a TL2hop-based query algorithm (TLQA) is designed by integrating an early termination strat-egy that reduces the communication overhead and boosts the query performance. Experimental results over eight data graphs demonstrate that the index time of BoundG is comparable to that of the state-of-the-art, and TL2hop significantly outperforms the state-of-the-art technique in terms of query response time (up to 4 orders of magnitude speedup).

Task assignment is a crucial issue in spatial crowd-sourcing. In most existing studies, the results of the task assignment cannot satisfy the workers and tasks at the same time. This is because only one-sided preferences are taken into account. Moreover, tasks are always assigned based on the locations of workers instead of the trajectories. Accordingly, they are not appropriate to the specific applications, such as carpool. Inspired by this, we investigate an interesting problem of task assignment, namely bilateral preference-aware task assignment (BPTA), with the goal of maximizing the overall satisfaction of workers and tasks by assigning tasks to suitable workers based on their routine trajectories. To tackle this problem effectively, we first propose greedy algorithms, namely task preference priority greedy and worker preference priority greedy algorithms, which are task-driven and worker-driven, respectively. Although these algorithms can solve the BPTA problem effectively, they cannot ensure the stability of the task assignment results. In other words, there can be better choices for some workers and tasks. Accordingly, we further explore deferred acceptance algorithms to find a stable matching for workers and tasks by simultaneously considering the preferences of workers and tasks. Moreover, two optimizing strategies, including a parallel strategy and a top-$k$strategy, are introduced to boost the performance in handling the BPTA problem. Extensive experiments on both real and synthetic datasets have validated the efficiency and effectiveness of our proposed algorithms.

Although multistage tasks involving multiple sequential goals are common in real-world applications, they are not fully studied in multi-agent reinforcement learning (MARL). To accomplish a multi-stage task, agents have to achieve cooperation on different subtasks. Exploring the collaborative patterns of different subtasks and the sequence of completing the subtasks leads to an explosion in the search space, which poses great challenges to policy learning. Existing works designed for single-stage tasks where agents learn to cooperate only once usually suffer from low sample efficiency in multi-stage tasks as agents explore aimlessly. Inspired by human’s improving cooperation through goal consistency, we propose Multi-Agent Goal Consistency (MAGIC) framework to improve sample efficiency for learning in multi-stage tasks. MAGIC adopts a goal-oriented actor-critic model to learn both local and global views of goal cognition, which helps agents understand the task at the goal level so that they can conduct targeted exploration accordingly. Moreover, to improve exploration efficiency, MAGIC employs two-level goal consistency training to drive agents to formulate a consistent goal cognition. Experimental results show that MAGIC significantly improves sample efficiency and facilitates cooperation among agents compared with state-of-art MARL algorithms in several challenging multistage tasks.

Oblivious RAM (ORAM) is a provable secure primitive to prevent access pattern leakage on the memory bus. By randomly remapping the data blocks and accessing redundant blocks, ORAM prevents access pattern leakage through ob-fuscation. Byte-addressable non-volatile memory (NVM) is considered as the candidate for main memory due to its better scalability, competitive performance, and persistent data store. While there is much prior work focusing on improving ORAM's performance on the conventional DRAM-based memory system, when the memory technology shifts to use NVM, ensuring an efficient crash-consistent ORAM is needed for security, correctness, and performance. Directly using traditional software-based crash consistency support for ORAM system is not only expensive but also insecure.

Network-wide monitoring is important for many network functions. However, monitoring data are often incomplete due to the need of sampling to reduce high measurement cost, system failure, and unavoidable transmission loss under severe communication. Instead of only targeting to estimate all missing monitoring data entries with a small set of measurement samples, we study a new order-preserved monitoring data estimation problem to accurately estimate the missing data entries while preserving the data entries’ order in the dataset. We propose a novel order-preserved tensor completion model that integrates both the low rank property and the order information into a joint learning problem to estimate the missing data. With well designed non-convex function to directly approximate the tensor rank and order-preserved constraint under the linear self-recovery method, our model can not only more accurately capture the low-rank property of monitoring data to increase the estimation performance of missing data, but also can capture the order information in monitoring data to ensure the estimation accuracy. Extensive experiments using four real datasets demonstrate that compared with the state-of-the-art tensor completion algorithms, our proposed algorithm can provide more accurate estimation and keep the value order of recovered entries to more effectively retrieve top-k large entries.

The pursuit of power-efficiency is popularizing asymmetric multicore processors (AMP) such as ARM big.LITTLE, Apple M1 and recent Intel Alder Lake with big and little cores. However, we find that existing scalable locks fail to scale on AMP and cause collapses in either throughput or latency, or both, because their implicit assumption of symmetric cores no longer holds. To address this issue, we propose the first asymmetry-aware scalable lock named LibASL. LibASL provides a new lock ordering guided by applications' latency requirements, which allows big cores to reorder with little cores for higher throughput under the condition of preserving applications' latency requirements. Using LibASL only requires linking the applications with it and, if latency-critical, inserting few lines of code to annotate the coarse-grained latency requirement. We evaluate LibASL in various benchmarks including five popular databases on Apple M1. Evaluation results show that LibASL can improve the throughput by up to 5 times while precisely preserving the tail latency designated by applications.

Deep learning services are extensively required and have powerful expected effects in a wide range of applications, such as auto-self driving, voice assistant and so on.Traditionally, deep learning services are mainly provided based on cloud computing, referred to as cloud intelligence services, where the deep learning model is deployed in the cloud, and endusers need to upload data through wireless and core network when requesting training and inference services. However, the cloud computing-based deep learning services have deficiencies in latency, privacy, etc. For example, the deep learning service users do not want their private data to leak, and the privacy problem is difficult to solve when the data is uploading to the cloud. With the widespread use of the Internet of Things (IoT) and the rapid development of mobile devices such as smartphones and IoT sensors, large amounts of data need to be used for a variety of real-time deep learning services, such as target recognition and voice recognition for smart cities, smart medical care, and the Internet of Vehicles (IoVs). With a certain network bandwidth, a large amount of data uploaded to the cloud will cause network congestion and greatly increase the response time. To meet the requirements of low latency, researchers have begun to consider the deployment of deep learning services in edges, i.e., edge intelligence service.In edge intelligence services, the computation capability and memory of processors (or devices) are different from a large. At the same time, the requirement of memory size of deep neural network (DNN) models is increasing, such as the memory usage for Alexnet and Resnet are 2.12G and 16.20G separately. Also, in DNN model design, the branches are becoming common, which brings the parallelism. Deploying deep learning models on multiple processors can support the large-scale DNN models and the parallel implementation of DNN model, where the computation of a deep learning model can be conducted in parallel is a possible solution to improve the efficiency of edge intelligence services. The key point in edge intelligence services is how to partition and assign the implementation of the DNN model.In this paper, we propose a novel latency-driven deep learning model placement method for efficient edge intelligence service. Model placement contains two procedures: model partition and sub-models assignment. In our method, we first convert a DNN model into an execution graph, which is a directed acyclic graph (DAG), and propose a novel latency-driven multilevel graph partition for the model. Then the partitioned submodels are heuristically assigned to available processors. To the best of our knowledge, it is the first work that proposes latency-driven graph partition algorithms for model placement. Extensive experiments on several commonly used DNN models and synthetic datasets show that our method can achieve the lowest execution latency with low complexity compared with other state-of-the-art model placement methods.

Abstract Nowadays, a new paradigm named mobile edge computing (MEC) is capable of supplying some cloud‐like functions at the edges of wireless networks, which enables vehicles to offload the computation intensive tasks on MEC servers with low latency. However, new challenges posed by the complex network environment and the mobility of vehicles are usually not covered by traditional offloading schemes. To solve such problems, we propose a heuristic task migration computation offloading (TMCO) scheme. Compared with traditional ones, TMCO can dynamically choose suitable places to offload the tasks for moving vehicles within deadline. For this purpose, the mobility of vehicle and strict delay deadline are considered comprehensively. We use hash table to store the number of tasks on the corresponding server and use random function to simulate the probability of task offloading. In terms of latency, experimental results suggest that the performance of TMCO is on average 10% higher than that of traditional full offloading schemes.

Mobile edge computing (MEC) is an emerging computational paradigm for providing storage and computing capabilities in network edge, to improve the experience of users, to shorten the delay, and to reduce the energy consumption of mobile devices. In this article, we consider a multiuser and multiserver scenario, where each user has an application composed of multiple independent tasks that need to be executed, and each MEC server is equipped on a base station (BS) for assisting mobile users to execute computation-intensive and time-sensitive tasks. Multiobjective optimization for joint task offloading, power assignment, and resource allocation is studied to maximize the offloading gains of users. A multivariable and multiobjective optimization problem with three objectives is constructed. An efficient multiobjective evolutionary algorithm is developed to solve the problems of minimizing the response time, minimizing the energy consumption, and minimizing the cost. Simulation results verify the effectiveness of our algorithm, and show the method significantly improves the user’s offloading benefits. According to the author’s knowledge, this is the first paper on the exploration of multiobjective optimization of multiuser with multiple tasks and multiserver MEC system, in which the worst user offloading revenue is regarded as the optimization objectives.

Mobile crowdsensing (MCS) has become a new paradigm for the environment sensing. However, the sparse sensory data prevent the practical and large-scale deployment of MCS systems. Recent studies have demonstrated that the matrix factorization is an effective technique which can estimate the missing sensory data entries based on a small set of observed data entries. However, there could be multiple sensory data sets with each regarded as a different view on the environment. Applying current matrix factorization individually to each data set, the recovery performance will be low as some data sets do not have enough observed data entries thus enough information. By partitioning the parameters involved in matrix factorization, we design some novel regularizations to encode the similarities among different data sets and specific knowledge in the single data set. Based on the regularizations, we propose one basic multiview matrix factorization (MVMF) model and one neural MVMF (NMVMF) model to combine multiple sensory data sets to mutually reinforce the estimation of each single data set. The extensive experimental results demonstrate that, with the help of other data sets, our models can estimate the missing entries in the data set with a very low sampling ratio accurately while the other five baseline algorithms cannot.

In the Internet of Everything era, various Internet of Things (IoT) devices have become popular, and the number of computing-intensive applications has increased substantially. As an emerging technology, mobile-edge computing (MEC) gives network edge nodes stronger computing and storage capabilities, bringing users a good Quality of Experience (QoE). By offloading some computing tasks to the edge for processing, the burden on IoT devices can be effectively reduced. However, this approach exacerbates the computing and storage resource depletion of the MEC server and the bandwidth and transmission cost of the wireless link used to offload computing tasks. Additionally, making an offloading decision online without future system status information is a considerable challenge. Therefore, we should study and design a reasonable offloading strategy to reduce the additional overhead, which is of significance. We establish a virtual queue model to describe the workload offloading problem of IoT devices in a two-layer MEC network. This is a stochastic optimization problem. Based on Lyapunov optimization, we transform the research problem into a deterministic optimization problem. A Lyapunov online energy consumption optimization algorithm (LOECOA) is proposed to effectively balance the system’s queue backlog and energy consumption. Based on theoretical analysis and a large number of experimental and numerical results, our algorithm performs better on energy consumption while satisfying the system constraints under a dynamic task arrival rate.

For the rapid growth computation requirements in big data and artificial intelligence area, CPU-GPU heterogeneous clusters can provide more powerful computing capacity compared to CPU clusters. The number of GPUs on single computing node is scalable, which greatly improves the computing capacity of the cluster under the condition of limited cluster size. However, there is a lack of the effective load-balancing scheduling model in multi-GPU hardware environment. This paper proposes AEML, an acceleration engine for multi-GPU load-balancing in distributed heterogeneous environment. AEML can effectively integrate GPUs into distributed processing framework and achieve great load-balance among multiple heterogeneous GPUs. We propose a heterogeneous task execution model based on multiple GPUs and multiple streams (MGMS), which can effectively balance the workload of multiple GPUs. MGMS model utilizes four core techniques: a fine-grained task mapping mechanism, a device resource unified management scheme, a novel resource-aware GPU task scheduling strategy and a feedback-based streams adjustment scheme. The implementation of AEML system is based on Spark 2.4.1 and NVIDIA CUDA 10.0. We comprehensively evaluate the performance of AEML with multiple typical benchmarks. Experimental results show that AEML can fully exploit the computing power of GPUs and achieve great load-balance among multiple heterogeneous GPUs.

Recent graph neural network (GNN) based methods for few-shot learning (FSL) represent the samples of interest as a fully-connected graph and conduct reasoning on the nodes flatly, which ignores the hierarchical correlations among nodes. However, real-world categories may have hierarchical structures, and for FSL, it is important to extract the distinguishing features of the categories from individual samples. To explore this, we propose a novel hierarchical graph neural network (HGNN) for FSL, which consists of three parts, i.e., bottom-up reasoning, top-down reasoning, and skip connections, to enable the efficient learning of multi-level relationships. For the bottom-up reasoning, we design intra-class k-nearest neighbor pooling (intra-class knnPool) and inter-class knnPool layers, to conduct hierarchical learning for both the intra- and inter-class nodes. For the top-down reasoning, we propose to utilize graph unpooling (gUnpool) layers to restore the down-sampled graph into its original size. Skip connections are proposed to fuse multi-level features for the final node classification. The parameters of HGNN are learned by episodic training with the signal of node losses, which aims to train a well-generalizable model for recognizing unseen classes with few labeled data. Experimental results on benchmark datasets have demonstrated that HGNN outperforms other state-of-the-art GNN based methods significantly, for both transductive and non-transductive FSL tasks. The dataset as well as the source code can be downloaded online1

Face aging has attracted widespread attention in recent years, but most studies are based on the same emotional situation. Is the same person’s aging in different emotional situations the same? To solve the above confusion, this paper proposes a novel face aging model DEF-Net, which consists of two parts: different emotional learnings (Emotion-Net) and face aging (Age-Net). Given a target emotion category, DEF-Net first assists the image from the original dataset to learn the emotion features through Emotion-Net and the generated dataset is used as the inputs of Age-Net. At the same time, multiple loss functions are used to ensure that the crucial information of the original image is not lost. Secondly, Age-Net, which has been pre-trained on the original dataset, began to adopt the generated dataset to learn the aging distribution under different emotions. Designed loss functions are utilized to ensure that the realistic target images generated by Age-Net do not lose the learned emotional characteristics. Finally, extensive experiments are used to verify the performance of DEF-Net. Compared with other state-of-the-art methods: (1) DEF-Net can learn different facial emotions across different datasets and generate corresponding realistic aging images; (2) the results achieved by our DEF-Net are demonstrated to be better than those by the model that performs face aging first and then learns different emotional characteristics.

Recognition of thyroid glands and tissues of the neck is vital for screening related diseases in ultrasound videos. This task is subjective, challenging, and dependent on the experience of sonographer in current clinical practice. The purpose is to develop a fully automated thyroid gland and tissues of neck recognition framework to assist doctors in distinguishing the boundaries of different tissues. In this paper, we propose a novel deep learning framework that consists of a feature extraction network, region proposal network, object detection head, and spatial pyramid RoIAlign-based segmentation head. Designed spatial pyramid RoIAlign can efficiently capture local and global context features, and aggregates the multiple context information that makes the result much more reliable. A large dataset is constructed to train the proposed method. The performance is evaluated using the COCO metrics. The experimental results demonstrate that the proposed deep learning method can effectively realize the automatic recognition of the thyroid gland and tissues of neck in ultrasound videos. Considering the clinical practical application scenarios, we developed an automatic recognition system of thyroid and neck tissue based on edge computing, which can expediently assist doctors in distinguishing the boundaries between different tissues.

Generalized sparse nonnegative matrix factorization (SNMF) has been proven useful in extracting information and representing sparse data with various types of probabilistic distributions from industrial applications, e.g., recommender systems and social networks.However, current solution approaches for generalized SNMF are based on the manipulation of whole sparse matrices and factor matrices, which will result in large-scale intermediate data.Thus, these approaches cannot describe the high-dimensional and sparse matrices in mainstream industrial and big data platforms, e.g., graphics processing unit (GPU) and multi-GPU, in an online and scalable manner. To overcome these issues, an online, scalable, and single-thread-based SNMF for CUDA parallelization on GPU (CUSNMF) and multi-GPU (MCUSNMF) is proposed in this article. First, theoretical derivation is conducted, which demonstrates that the CUSNMF depends only on the products and sums of the involved feature tuples. Next, the compactness, which can follow the sparsity pattern of sparse matrices, endows the CUSNMF with online learning capability and the fine granularity gives it high parallelization potential on GPU and multi-GPU. Finally, the performance results on several real industrial datasets demonstrate the linear scalability of the time overhead and the space requirement and the validity of the extension to online learning. Moreover, CUSNMF obtains speedup of 7X on a P100 GPU compared to that of the state-of-the-art parallel approaches on a shared memory platform.

With the popularity of 3-D sensors in industrial Internet of Things (IIoT), point clouds learning is increasingly important. In this article, we propose a novel multilevel attention based U-shape graph neural network (MAUGNN) for point clouds learning, which can effectively learn the features from low-level to high-level and fuse multiple-level features based on the graph neural networks and attention mechanism. There are three parts in MAUGNN: encoder, decoder, and connections. In the encoder and decoder, we design an attention-based graph convolution to explore the structural information for point clouds. During the encoder, a structure-aware attention pooling is proposed to support down-sampling on point cloud data. To adaptively fuse coarse-grained features from the encoder and fine-grained features from the decoder together, we also propose a structure-aware attention skip connection mechanism. Extensive experiments on popular point cloud datasets demonstrate the superior performance of our MAUGNN over state-of-the-art baselines.

The apical four-chamber (A4C) view in fetal echocardiography is a prenatal examination widely used for the early diagnosis of congenital heart disease (CHD). Accurate segmentation of A4C key anatomical structures is the basis for automatic measurement of growth parameters and necessary disease diagnosis. However, due to the ultrasound imaging arising from artefacts and scattered noise, the variability of anatomical structures in different gestational weeks, and the discontinuity of anatomical structure boundaries, accurately segmenting the fetal heart organ in the A4C view is a very challenging task. To this end, we propose to combine an explicit Feature Pyramid Network (FPN), MobileNet and UNet, i.e., MobileUNet-FPN, for the segmentation of 13 key heart structures. To our knowledge, this is the first AI-based method that can segment so many anatomical structures in fetal A4C view. We split the MobileNet backbone network into four stages and use the features of these four phases as the encoder and the upsampling operation as the decoder. We build an explicit FPN network to enhance multi-scale semantic information and ultimately generate segmentation masks of key anatomical structures. In addition, we design a multi-level edge computing system and deploy the distributed edge nodes in different hospitals and city servers, respectively. Then, we train the MobileUNet-FPN model in parallel at each edge node to effectively reduce the network communication overhead. Extensive experiments are conducted and the results show the superior performance of the proposed model on the fetal A4C and femoral-length images.

Defect detection of trains is of great significance for operation safety and maintenance efficiency for railway maintenance. Nowadays, China railway system utilizes high-speed line scan cameras to capture images of critical parts of moving trains. The visual inspection on the images still heavily relies on manual interpretation. To reduce the labor requirements, we propose a novel two-stage deep learning based framework for component defect detection of moving trains. The proposed framework is composed of two major successive stages: detecting train components by using our proposed hierarchical object detection scheme (HOD), and detecting component defects based on multiple neural networks and image processing methods. Our proposed HOD can effectively detect and localize train components from large to small in a hierarchical way. Furthermore, a gated feature fusion method that can extract and combine the hierarchical contextual features and spatial contexts is also proposed to improve the performance. To the best of our knowledge, it is the first time in the literature that component defect detection of moving trains is systematically analyzed. Extensive experiments on real images from China railway system have demonstrated that our framework outperforms the state-of-the-art baselines significantly.

In Vehicular Ad-hoc Networks (VANETs), privacy protection and data security during network transmission and data analysis have attracted attention. In this paper, we apply deep learning, blockchain, and fully homomorphic encryption (FHE) technologies in VANETs and propose a Decentralized Privacy-preserving Deep Learning (DPDL) model. We propose a Decentralized VANETs (DVANETs) architecture, where computing tasks are decomposed from centralized cloud services to edge computing (EC) nodes, thereby effectively reducing network communication overhead and congestion delay. We use blockchain to establish a secure and trusted data communication mechanism among vehicles, roadside units, and EC nodes. In addition, we propose a DPDL model to provide privacy-preserving data analysis for DVANET, where the FHE algorithm is used to encrypt the transportation data on each EC node and input it into the local DPDL models, thereby effectively protecting the privacy and credibility of vehicles. Moreover, we further use blockchain to provide a decentralized and trusted DPDL model update mechanism, where the parameters of each local DPDL model are stored in the blockchain for sharing with other distributed models. In this way, all distributed models can update their models in a credible and asynchronous manner, avoiding possible threats and attacks. Extensive simulations are conducted to evaluate the effectiveness, practicality, and robustness of the proposed DVANET system and DPDL models.

Point cloud, an efficient 3D object representation, plays an indispensable role in autonomous driving technologies, such as object avoidance, localization, and map building. The analysis of point clouds (e.g., 3D segmentation) is essential to exploit the informative value of point clouds for such applications. The main challenge remains to effectively and completely extract high-level point cloud feature representations. To this end, we present a novel multi-task Y-shaped graph neural network to explore 3D point clouds, referred to as MTYGNN. By extending the conventional U-Net, MTYGNN contains two main branches to simultaneously perform classification and segmentation tasks in point clouds. Meanwhile, the classification prediction is fused together with the semantic features as the scene context to make the segmentation task more accurate. Furthermore, we consider the homoscedastic uncertainty of each task to calculate the weights of multiple loss functions to ensure that tasks do not negatively interfere with each other. The proposed MTYGNN is evaluated on popular point cloud datasets in traffic scenarios. Experimental results demonstrate that our framework outperforms the state-of-the-art baseline methods.

Time-series forecasting is an important problem across a wide range of domains. Designing accurate and prompt forecasting algorithms is a non-trivial task, as temporal data that arise in real applications often involve both non-linear dynamics and linear dependencies, and always have some mixtures of sequential and periodic patterns, such as daily, weekly repetitions, and so on. At this point, however, most recent deep models often use Recurrent Neural Networks (RNNs) to capture these temporal patterns, which is hard to parallelize and not fast enough for real-world applications especially when a huge amount of user requests are coming. Recently, CNNs have demonstrated significant advantages for sequence modeling tasks over the de-facto RNNs, while providing high computational efficiency due to the inherent parallelism. In this work, we propose HyDCNN, a novel hybrid framework based on fully Dilated CNN for time-series forecasting tasks. The core component in HyDCNN is a proposed hybrid module, in which our proposed position-aware dilated CNNs are utilized to capture the sequential non-linear dynamics and an autoregressive model is leveraged to capture the sequential linear dependencies. To further capture the periodic temporal patterns, a novel hop scheme is introduced in the hybrid module. HyDCNN is then composed of multiple hybrid modules to capture the sequential and periodic patterns. Each of these hybrid modules targets on either the sequential pattern or one kind of periodic patterns. Extensive experiments on five real-world datasets have shown that the proposed HyDCNN is better compared with state-of-the-art baselines and is at least 200% better than RNN baselines. The datasets and source code will be published in Github to facilitate more future work.

Multi-set membership query is a fundamental issue for network functions such as packet processing and state machines monitoring. Given the rigid query speed and memory requirements, it would be promising if a multi-set query algorithm can be designed based on Bloom filter (BF), a space-efficient probabilistic data structure. However, existing efforts on multi-set query based on BF suffer from at least one of the following drawbacks: low query speed, low query accuracy, limitation in only supporting insertion and query operations, or limitation in the set size. To address the issues, we design a novel B h sequence-based Bloom filter (B h BF) for multi-set query, which supports four operations: insertion, query, deletion, and update. In B h BF, the set ID is encoded as a code in a B h sequence. Exploiting good properties of B h sequences, we can correctly decode the BF cells to obtain the set IDs even when the number of hash collisions is high, which brings high query accuracy. In B h BF, we propose two strategies to further speed up the query speed and increase the query accuracy. On the theoretical side, we analyze the false positive and classification failure rate of our B h BF. Our results from extensive experiments over two real datasets demonstrate that B h BF significantly advances state-of-the-art multi-set query algorithms.

Recently, deep learning-based approaches have achieved superior performance on object detection applications. However, object detection for industrial scenarios, where the objects may also have some structures and the structured patterns are normally presented in a hierarchical way, is not well investigated yet. In this work, we propose a novel deep learning-based method, hierarchical graphical reasoning (HGR), which utilizes the hierarchical structures of trains for train component detection. HGR contains multiple graphical reasoning branches, each of which is utilized to conduct graphical reasoning for one cluster of train components based on their sizes. In each branch, the visual appearances and structures of train components are considered jointly with our proposed novel densely connected dual-gated recurrent units (Dense-DGRUs). To the best of our knowledge, HGR is the first kind of framework that explores hierarchical structures among objects for object detection. We have collected a data set of 1130 images captured from moving trains, in which 17 334 train components are manually annotated with bounding boxes. Based on this data set, we carry out extensive experiments that have demonstrated our proposed HGR outperforms the existing state-of-the-art baselines significantly. The data set and the source code can be downloaded online at https://github.com/ChengZY/HGR.

For solving dynamic generalized Lyapunov equation, two robust finite-time zeroing neural network (RFTZNN) models with stationary and nonstationary parameters are generated through the usage of an improved sign-bi-power (SBP) activation function (AF). Taking differential errors and model implementation errors into account, two corresponding perturbed RFTZNN models are derived to facilitate the analyses of robustness on the two RFTZNN models. Theoretical analysis gives the quantitatively estimated upper bounds for the convergence time (UBs-CT) of the two derived models, implying a superiority of the convergence that varying parameter RFTZNN (VP-RFTZNN) possesses over the fixed parameter RFTZNN (FP-RFTZNN). When the coefficient matrices and perturbation matrices are uniformly bounded, residual error of FP-RFTZNN is bounded, whereas that of VP-RFTZNN monotonically decreases at a super-exponential rate after a finite time, and eventually converges to 0. When these matrices are bounded but not uniform, residual error of FP-RFTZNN is no longer bounded, but that of VP-RFTZNN still converges. These superiorities of VP-RFTZNN are illustrated by abundant comparative experiments, and its application value is further proved by an application to robot.

Predicting the popularity of web contents in online social networks is essential for many applications. However, existing works are usually under non-incremental settings. In other words, they have to rebuild models from scratch when new data occurs, which are inefficient in big data environments. It leads to an urgent need for incremental prediction, which can update previous results with new data and conduct prediction incrementally. Moreover, the promising direction of group-level popularity prediction has not been well treated, which explores fine-grained information while keeping a low cost. To this end, we identify the problem of incremental group-level popularity prediction, and propose a novel model IGPP to address it. We first predict the group-level popularity incrementally by exploiting the incremental CANDECOMP/PARAFCAC (CP) tensor decomposition algorithm. Then, to reduce the cumulative error by incremental prediction, we propose three strategies to restart the CP decomposition. To the best of our knowledge, this is the first work that identifies and solves the problem of incremental group-level popularity prediction. Extensive experimental results show significant improvements of the IGPP method over other works both in the prediction accuracy and the efficiency.

Deep learning models are widely used in daily life, which bring great convenience to our lives, but they are vulnerable to attacks. How to build an attack system with strong generalization ability to test the robustness of deep learning systems is a hot issue in current research, among which the research on black-box attacks is extremely challenging. Most current research on black-box attacks assumes that the input dataset is known. However, in fact, it is difficult for us to obtain detailed information for those datasets. In order to solve the above challenges, we propose a multi-sample generation model for black-box model attacks, called MsGM. MsGM is mainly composed of three parts: multi-sample generation, substitute model training, and adversarial sample generation and attack. Firstly, we design a multi-task generation model to learn the distribution of the original dataset. The model first converts an arbitrary signal of a certain distribution into the shared features of the original dataset through deconvolution operations, and then according to different input conditions, multiple identical sub-networks generate the corresponding targeted samples. Secondly, the generated sample features achieve different outputs through querying the black-box model and training the substitute model, which are used to construct different loss functions to optimize and update the generator and substitute model. Finally, some common white-box attack methods are used to attack the substitute model to generate corresponding adversarial samples, which are utilized to attack the black-box model. We conducted a large number of experiments on the MNIST and CIFAR-10 datasets. The experimental results show that under the same settings and attack algorithms, MsGM achieves better performance than the based models.

This article proposes a pattern-based prefetching scheme with the support of adaptive cache management, at the flash translation layer of solid-state drives ( SSDs ). It works inside of SSDs and has features of OS dependence and uses transparency. Specifically, it first mines frequent block access patterns that reflect the correlation among the occurred I/O requests. Then, it compares the requests in the current time window with the identified patterns to direct prefetching data into the cache of SSDs. More importantly, to maximize the cache use efficiency, we build a mathematical model to adaptively determine the cache partition on the basis of I/O workload characteristics, for separately buffering the prefetched data and the written data. Experimental results show that our proposal can yield improvements on average read latency by 1.8 %– 36.5 % without noticeably increasing the write latency, in contrast to conventional SSD-inside prefetching schemes.

Integrating user ends (UEs), edge servers (ESs), and the cloud into end-edge-cloud computing (EECC) can enhance the utilization of resources and improve quality of experience (QoE). However, the performance of EECC is significantly affected by its architecture. In this article, we classify EECC into two computing architectures types according to the visibility and accessibility of the cloud to UEs, i.e., hierarchical end-edge-cloud computing (Hi-EECC) and horizontal end-edge-cloud computing (Ho-EECC). In Hi-EECC, UEs can offload their tasks only to ESs. When the resources of ESs are exhausted, the ESs request the cloud to provide resources to UEs. In Ho-EECC, UEs can offload their tasks directly to ESs and the cloud. In this article, we construct a potential game for the EECC environment, in which each UE selfishly minimizes its payoff, study the computation offloading strategy optimization problems, and develop two potential game-based algorithms in Hi-EECC and Ho-EECC. Extensive experiments with real-world data are conducted to demonstrate the performance of the proposed algorithms. Moreover, the scalability and applicability of the two computing architectures are comprehensively analyzed. The conclusions of our work can provide useful suggestions for choosing specific computing architectures under different application environments to improve the performance of EECC and QoE.

Computing resource configuration and site selection of edge servers (ESs) are two critical steps to build up a mobile edge computing (MEC) platform. In this paper, the joint optimization problem of configuration and placement for ES in the MEC environment is investigated. First, we treat each ES as an M/G/m queueing model, and establish mathematical models to characterize the MEC environment, such that the performance and operational expenditures (OPEX) of the system can be calculated analytically. Then, we design a two-stage method and develop a series of algorithms based on bisection algorithm and genetic algorithm (GA) to obtain the optimal configuration scheme and sub-optimal placement scheme (including the deployment quantity) of ESs, with the goal of minimizing OPEX while maintaining system performance at a predetermined level. Finally, we conduct experiments based on a real base station dataset provided by Shanghai Telecom to show the effectiveness of the proposed algorithms. To the best of our knowledge, this work is the first research of the joint optimization problem of configuration and placement for ES in the MEC environment, where the main objective is to increase the cost efficiency.

Recently, with the large-scale outbreak of the global financial crisis and public safety incidents (such as COVID-19), high-performance computing has been widely applied to risk prediction, vaccine development, and other fields. In scenarios where high-performance computing infrastructure responds to the instantaneous explosion of computing demands, a crucial issue is to provide large-scale flexible allocation and adjustment of computing capability by rapidly constructing computing clusters. Existing large-scale computing cluster deployment solutions usually utilize source code deployment or other deployment tools. The great challenge of existing deployment methods is to reduce excessive image distribution time and refrain from configuration defects. In this article, we design an intelligent distributed registry deployment (IDRD) architecture based on the OpenStack cloud platform, which adaptively places distributed image repositories using the containerized deployment of multiple registries. We propose a server load priority algorithm to solve multiple registries placement problems in IDRD. Furthermore, we devise a clustering algorithm based on demand density that can optimize the global performance of IDRD and improve large-scale cluster load balancing capabilities, which has been implemented in the TianHe Supercomputing environment. Extensive experimental results demonstrate that IDRD can effectively reduce$30\%$-$50\%$of the distribution time of component images and significantly improve the efficiency of large-scale cluster deployment.

Large-scale Internet applications running on data centers are typically instantiated as a set of containers. Assigning a container to its affinity machine can reduce communication and transport costs while assigning it to the anti-affinity machine may affect the proper operation of the container. Existing container scheduling methods cannot accommodate these two types of requirements. In order to reduce the operation and maintenance cost of data centers, this paper focuses on the container instance allocation problem in heterogeneous server cluster, and proposes a global cost-aware scheduling algorithm (GCCS) to solve it. The purpose is to minimize the total power consumption of the cluster from a global perspective, while trying to meet the affinity/anti-affinity requirements of applications. We study the number of containers per server selected by the application, model it as an integer linear program (ILP), and then propose a heuristic search algorithm to repair the relaxation solution of the ILP into a suboptimal feasible solution. In particular, we use Bayesian optimizer to perform a number of automated development and exploration processes for the selection of the cost coefficient. The experiments are carried out with the best cost coefficient recommended by Bayesian optimizer. Finally, the results demonstrate that GCCS can significantly reduce the total power consumption of the cluster, while maintaining a high affinity satisfaction ratio.

Monodirectional tissue P systems with promoters are natural inspired parallel computing paradigms, where only symport rules are permitted, and with the restriction of “monodirectionality”, objects for two given regions are transferred in one direction. In this article, a novel kind of P systems, monodirectional evolutional symport tissue P systems with promoters (MESTP P systems) is raised, where objects may be revised during the movement between two regions. The computational theory of MESTP P systems that rules are employed in a flat maximally parallel pattern is investigated. We prove that finite natural number sets are created by MESTP P systems applying one cell, at most 1 promoter and all evolutional symport rules having a maximal length 2 or with arbitrary number of cells, promoters and all evolutional symport rules having a maximal length 2. MESTP P systems are Turing universal when two cells, at most 1 promoter and all evolutional symport rules having a maximal length 2 are employed. In addition, with the help of cell division mechanism, monodirectional evolutional symport tissue P systems with promoters and cell division (MESTPD P systems) are employed to solve NP-complete (the SAT) problem, where system uses at most 1 promoter and all evolutional symport rules having a maximal length 3. These results show that MESTP(D) P systems are still computationally powerful even if monodirectionality control mechanism is imposed, thereby developing membrane algorithms for MESTP(D) P systems is theoretically possible as well as potentially exploitable.

As a fundamental problem in graph analysis, core decomposition aims to compute the core numbers of vertices in a given graph. It is a powerful tool for mining important graph structures. For dynamic graphs with real-time updates of vertices/edges, core maintenance has been utilized to update the core numbers of vertices. The previous approaches to core maintenance face challenges in terms of storage and efficiency. In this article, we investigate distributed approaches to core maintenance on a pregel-like system, which is a famous graph computing system. We first design a core decomposition algorithm to obtain core numbers of vertices in a given graph. Based on it, a distributed batch-stream combined algorithm (DBCA) is devised to efficiently maintain the core numbers when vertex/edge updates happen. In particular, we introduce a new task assignment strategy to DBCA based on diversity of the edge-cores of updated edges. To ensure that DBCA can accurately process core maintenance, we develop a message interaction protocol to resolve the problem of crosstalk among different tasks. Comprehensive experiments have been conducted on real/synthetic graphs, more specifically, in two typical distributed environments built on Supercomputing Center and Alibaba Cloud. The experiment results demonstrate that our proposed algorithms are efficient and scalable.

Transformer-based deep neural networks have recently swept the field of natural language processing due to their outstanding performance, and are gradually spreading to more applications such as image/video processing. However, compared with general DNNs, training a sizeable transformer-based model is further time-consuming and memory-hungry. The existing distributed training strategies for general DNNs are not appropriate or can not efficiently handle transformer-based networks. In view of this, we propose an intra-layer model parallelization optimization strategy, AccTFM, which introduces a novel fine-grained pipeline execution and hybrid communication compression strategy to overcome the synchronization bottleneck. Specifically, on one hand, it first decouples the inter-layer computation and communication dependencies, and then searches for the optimal partitioning strategy to maximize the overlap of computation and communication. On the other hand, the hybrid communication compression module consists of token-level top-$k$sparsification and piecewise quantization methods aiming at minimizing communication traffic. Experimental results show that AccTFM accelerates transformer-based DNNs training by up to 2.08x compared to state-of-the-art distributed training techniques.

Set reachability query in directed graphs has a plethora of graph-based applications such as dependency analysis and graph centrality calculation. Given two sets$S$and$T$of source and target vertices, set reachability query needs to acquire all pairs$(s,t)$where$s{\in }S$,$t{\in }T$and$s$can reach$t$. The state-of-the-art approach distributed set reachability (DSR) investigates the set reachability query in a distributed environment and adopts a static graph-based index to enhance the query efficiency. Nevertheless, DSR needs to store the graph-based index in all partitions, which causes a huge space overhead. Furthermore, it cannot efficiently solve the negative query$(s,t)$where$s$cannot reach$t$, since DSR needs to traverse the whole reachable paths and becomes unable to efficiently reduce the computations. To alleviate these issues, we propose a novel multi-level 2-hop (ML2hop) index for the set reachability query in a distributed environment. Based on ML2hop, we further present a bi-directional query algorithm, called MLQA, to achieve efficient support for both positive and negative queries in Pregel-like systems. Generally, MLQA is equipped with the following three significant properties: (1) Low computation costs. It reduces redundant local computations in each partition by controlling the rounds of path traversals. (2) Low communication costs. It restricts the message exchange among different partitions within one single round with guaranteed accuracy of query results. (3) High parallelism. It adopts a bi-directional query technique for message propagation, achieving the better query efficiency than the forward-traversal query strategy utilized in DSR. Experimental results over several real-world graphs demonstrate that MLQA significantly outperforms the state-of-the-art algorithm by up to two orders of magnitude speedup.

The high concurrency and high throughput characteristics of graphics processing units (GPUs) have made researchers continue to use it to optimize distributed parallel computing architectures. With the upgrading of processor architecture, GPUs allow multiple kernels to execute concurrently through stream queues. However, due to the different hardware characteristics and kernel properties in distributed architectures, existing research lacks careful consideration of optimization schemes for concurrent streams and kernel block sizes. Unreasonable stream concurrency and kernel block size configuration will lead to prolonged execution time and waste of computing resources during application execution. Therefore, we propose a multi-GPU multi-stream co-concurrency mechanism (MGSC) in a distributed heterogeneous environment, dynamically adjusting the number of concurrent streams and exploring the optimal block size in task scheduling. According to the memory resources and startup overhead occupied in concurrent stream scheduling, a resource-aware concurrent stream adaptive adjustment mechanism is proposed, which can dynamically adjust the number of streams. To explore the optimal block size, we abstract it as a multi-armed bandit problem (MAB) and propose a block size adjustment algorithm based on the upper confidence bound (UCB). We implement MGSC in Spark 3.1.1 and NVIDIA CUDA 11.2. We conduct comparative experiments with multiple typical benchmarks to evaluate the performance of MGSC. The experimental results show that the algorithm can make full use of the computing power of the GPU and significantly reduce the execution time of tasks.

The ongoing digitalization enables stable control processes and smooth operations of Industrial Control Systems (ICSs). A direct consequence of the highly interconnected architecture of ICSs is the introduced cyber vulnerability and increasing cyber security threats to ICSs. Numerous researches pay attention to the security problem of ICSs. However, most current researches face two challenges. First, the interaction problem between cyber layer and physical layer of ICSs may result incorrect attack response strategies. Second, ICSs are real-time systems, but existing defense decision algorithms based on game theory or reinforcement learning techniques have high computational complexity, which prevents it from making decisions quickly. In this paper, we design a new multi-attribute based reward quantitative method and propose a multi-attribute based Q-learning algorithm to resolve the interaction problem. In addition, to overcome the limitation of slow convergence, we develop an effective parallel Q-learning (PQL) algorithm to quickly find the optimal strategy. The experimental results show the effectiveness of the PQL algorithm. Compared with the Q-learning algorithm (QL) and the deep Q-network (DQN) algorithm, our proposed solution can reduce the average completion time by 12.5%-37%.

Deep learning services based on cloud computing have deficiencies in latency, privacy, etc. To meet the requirements of low latency, researchers have begun to consider the deployment of deep learning services in edges, i.e., edge intelligence service. Deploying deep learning models on multiple processors or devices so that the computation of a deep learning model can be conducted in parallel is a possible solution to improve the efficiency of edge intelligence services. In this article, we propose a novel latency-driven deep learning model placement method for efficient edge intelligence service. Model placement contains two procedures: model partition and sub-models assignment. In our method, we first convert the model into execution graphs and propose a novel latency-driven multilevel graph partition for the model. Then the partitioned sub-models are heuristically assigned to available processors. To the best of our knowledge, it is the first work that proposes latency-driven graph partition algorithms for model placement. Extensive experiments on several commonly used DNN (deep neural network) models and synthetic datasets show that our method can achieve the lowest execution latency with low complexity compared with other state-of-the-art model placement methods.

In this article, two Adams–Bashforth-type integration-enhanced discrete-time zeroing neural dynamic (ADTIZD) models are proposed to solve the time-varying complex Sylvester equation (TVCSE) problem in the first time. In ADTIZD models, Adams–Bashforth discrete formulas as novel discrete formulas are used, giving our ADTIZD models higher accuracy [truncation error being$O(\tau ^{5})$] but less time and space complexity than the ordinary multi-instant models. Enhanced by the integration part, the ADTIZD models can resist large additive noises, where even constant noises cannot decrease their precision. All convergence and robustness performance conclusions about our ADTIZD models are supported by rigorous theoretical proofs and numerical experiments. More comparisons between ADTIZD models and other discrete-time zeroing neural network models are shown in these experiments too. The efficacy of ADTIZD models is finally been validated in the simulation of adopting them in controlling a robotic manipulator.

Zeroing neural network (ZNN), an effective method for tracking solutions of dynamic equations, has been developed and improved by various strategies, typically the application of nonlinear activation functions (AFs) and varying parameters (VPs). Unlike VPs, AFs applied in ZNN models act directly on real-time error. The processing unit of v needs to obtain neural state in real time. In the implementation process, highly nonlinear AFs become an important cause of time delays, which eventually leads to instability and oscillation. However, most studies focus on exploring new theoretically valid AFs to improve performance of ZNNs, while ignoring the adverse effects of highly nonlinear AFs. The nonlinearity of AFs requires us fully consider time-delay tolerance of ZNNs using nonlinear AFs, so as to ensure that the model is not unstable even when disturbed by time delays. In this work, delay-perturbed generalized ZNN (DP-GZNN) is proposed to investigate time-delay tolerance of generalized ZNN (G-ZNN) in solving dynamic Lyapunov equation. Considering the nonlinearity of AFs, two delay terms are elegantly added to G-ZNN and DP-GZNN is then derived. After rigorous mathematical derivations, sufficient conditions in a linear matrix inequality (LMI) manner are presented for global convergence of DP-GZNN. Through rich numerical experiments, hyperparameters involved in the analysis process are discussed in detail. Comparative simulations are also conducted to compare the ability of different ZNN models to resist time delays. It is worth to mention that this is the first time to consider the ability of G-ZNN to resist discrete and distributed time delays.

Heterogeneous Computing Systems (HCSs) have developed rapidly due to their high performance and low cost, and have been adopted by more and more applications. Energy consumption, reliability, and schedule length are the core issues of HCSs. Due to the negative correlation between frequency and reliability, DVFS-supported HCSs requires high energy consumption and a long schedule length to obtain high reliability, which resulting in performance degradation. In this paper, we focus on the reliability and performance-aware scheduling for energy-constrained parallel applications on HCSs. First, we design an energy pre-allocation mechanism based on Energy Demand Rate (EDR) to pre-allocate energy reasonably. Second, we propose an EDR-aware Maximizing Reliability of Energy-Constrained parallel applications (EMREC) scheduling algorithm. Third, considering that maximize reliability will cause the schedule length to be too long and unacceptable, we further highlight the concept of Reliability Performance Ratio (RPR). Finally, we propose a Maximizing RPR with Energy-Constrained parallel applications (MRPEC) scheduling algorithm, which enables parallel applications have a smaller schedule length while with high reliability. Extensive experimental results in real-world and randomly generated applications show the effectiveness of the proposed algorithms under different conditions.

Virtualized networked datacenters (VNDCs) are gaining considerable attention for stochastic task execution under real-time constraints. However, the problem of efficiently minimizing the high energy consumption while ensuring high quality of service (QoS) in VNDCs has not been fully addressed. Although many solutions have been proposed to address this challenge, they are not efficient and only consider one or two of the energy consuming resources of VNDCs. To this end, an adaptive energy-aware algorithm,MCEC, that efficiently reduces the energy consumption of VNDCs while ensuring high QoS is proposed. Different from the existing approaches, the MCEC algorithm considers energy consumed by computing resources, virtual machine (VM) reconfiguration, communication resources and storage media resources while meeting user QoS requirements defined in the service level agreement (SLA). To validate the effectiveness of our algorithm, we carried out extensive experiments and compared the performance of our algorithm with existing baseline algorithms. The results of the experiments show that our algorithm substantially outperforms the baseline algorithms with respect to reducing energy consumption while respecting the service level agreement.

Recently, graph neural networks (GNNs) have achieved great success for graph representation learning tasks. Enlightened by the fact that numerous message passing redundancies exist in GNNs, we propose DyGNN, which speeds up GNNs by reducing redundancies. DyGNN is supported by an algorithm and architecture co-design. The proposed algorithm can dynamically prune vertices and edges during execution without accuracy loss. An architecture is designed to support dynamic pruning and transform it into performance improvement. DyGNN opens new directions for accelerating GNNs by pruning vertices and edges. DyGNN gains average $2\times$ speedup with accuracy improvement of 4% compared with state-of-the-art GNN accelerators.

Deep Neural Networks (DNNs) are widely investigated due to their striking performance in various applications of artificial intelligence. However, with DNNs becoming larger and deeper, the computing resource of a single hardware accelerator is insufficient to meet the training requirements of popular DNNs. Hence, it is required to train them using multiple accelerators in a distributed setting. For a better utilization of the accelerators and a faster training, it is necessary to partition the whole process into segments that can run in parallel. However, in this context, intra-layer parallelization techniques (i.e., data and model parallelization) often face communication and memory bottlenecks, while the performance and resource utilization of inter-layer parallelization techniques (i.e., using pipelining) depend on the partitioning possibilities of the model. We present EffTra, a synchronous hybrid parallelization strategy, that uses a combination of intra-layer and inter-layer parallelism to realize a distributed training of DNNs. EffTra employs the idea of dynamic programming to try to search for the optimal partitioning of a DNN model and assigns devices to the obtained partitions. Our evaluation shows that EffTra accelerates training by up to 2.0x and 1.78x compared to state-of-the-art inter-layer (i.e., GPipe) and intra-layer (i.e., data parallelism) parallelization techniques respectively.

Time series forecasting is very important and plays critical roles in many applications. However, making accurate forecasting is a challenge task due to the requirements of learning complex temporal and spatial patterns and combating noise during the feature learning. To address the challenge issues, we propose TEGMNet, a Tensor projection Embedding and Gated Memory Network for multivariate time series forecasting. To more accurately extract local features and reduce the influence of noise, we propose to amplify the data using several data transformation techniques based on MDT (Multi-way delay embedding transform) and TFNN (tensor factorized neural network) to transform the original 2D matrix data to low dimensional 3D tensor data. The local features are then extracted through convolution and LSTM upon the 3D tensor. We also design a long-term feature extraction module based on the structure of gated memory network, which can largely enhance the longterm pattern feature learning ability when the multivariate time series has complex long-term dependencies with dynamic-period patterns. We have done extensive experiments by comparing our TEGMNet with 7 baseline algorithms using 4 real data sets. The experiment results demonstrate that TEGMNet can achieve very good prediction performance even through the data are polluted with noise.

The current research on adversarial attacks aims at a single model while the research on attacking multiple models simultaneously is still challenging. In this paper, we propose a novel black-box attack method, referred to as MBbA, which can attack multiple black-boxes at the same time. By encoding input image and its target category into an associated space, each decoder seeks the appropriate attack areas from the image through the designed loss functions, and then generates effective adversarial examples. This process realizes end-to-end adversarial example generation without involving substitute models for the black-box scenario. On the other hand, adopting the adversarial examples generated by MBbA for adversarial training, the robustness of the attacked models are greatly improved. More importantly, those adversarial examples can achieve satisfactory attack performance, even if these black-box models are trained with the adversarial examples generated by other black-box attack methods, which show good transferability. Finally, extensive experiments show that compared with other state-of-the-art methods: (1) MBbA takes the least time to obtain the most effective attack effects in multi-black-box attack scenario. Furthermore, MBbA achieves the highest attack success rates in a single black-box attack scenario; (2) the adversarial examples generated by MBbA can effectively improve the robustness of the attacked models and exhibit good transferability.

Task scheduling in cloud environments is the problem of assigning and executing computational tasks on the available cloud resources. Effective task scheduling can improve processor utilization, reduce processor energy consumption, and improve user experience. Large-scale task scheduling under multiple constraints is an NP-complete problem. The traditional task scheduling algorithm cannot be applied to large-scale scheduling, either because of high time complexity or because its heuristic algorithm cannot be applied to complexly large-scale scenarios. The problem of large-scale task scheduling is gradually becoming a challenge in cloud computing. The article proposes a topology-based multilevel algorithm for large-scale task scheduling in clouds. Based on the topological order of the graph, multi-level coarsening is performed on the large-scale graphs, and then uses the traditional scheduling algorithm for the initial scheduling of the coarse graph, and then refine the initial scheduling result during its uncoarsen phrase. It can perform fast and efficient scheduling of large-scale task graphs. At the same time, it has good compatibility, which can be combined with excellent traditional scheduling algorithms.

Functional verification of large scale circuit design is a basic problem in Very Large Scale Integrated (VLSI) design. With the increasing scale of the circuit, it is urgent to divide the whole large scale circuit into some smaller sub-circuits so as to perform parallel functional verification on multiple hardware processors. The partition problem of hardware-accelerated functional verification can be regarded as a graph partition problem. However, unlike the traditional graph partition requirements for minimum cutting, the hardware-accelerated functional verification partition needs to reduce the simulation depth and improve the parallelism of the simulation. Therefore, partition for hardware-accelerated functional verification is a problem combined with graph partitioning and schedule. While the traditional schedule algorithms have high complexity and cannot handle large scale Directied Acyclic Graph (DAG) scheduling. To tackle the parallelism, depth, and cut edge problem, we design a new method, called path-metis. Path-metis combines the scheduling idea, such as the critical path information and task priority of the DAG, into the traditional multilevel partitioning method. Our preliminary experiments on real circuits show the effectiveness of the method, and the simulation depth can be reduced by about 11.35% on average compared with metis only with 27.58% cut size increasing.

Multifrontal QR algorithm, which consists of symbolic analysis and numerical factorization, is a high-performance algorithm for orthogonal factorizing sparse matrix. In this work, a graph convolutional network (GCN) for adaptively selecting the optimal reordering algorithm is proposed in symbolic analysis. Using our GCN adaptive classifier, the average numerical factorization time is reduced by 20.78% compared with the default approach, and the additional memory overhead is approximately 4% higher than that of prior work. Moreover, for numerical factorization, an optimized tasks stream parallel processing strategy is proposed and a more efficient computing task mapping framework for NUMA architecture is adopted in this paper, which called STM-Multifrontal QR factorization. Numerical experiments on the TaiShan Server show average 1.22x performance gains over the original SuiteSparseQR. Nearly 80% of datasets have achieved better performance compared with the MKL sparse QR on Intel Xeon 6248.

Recent studies have demonstrated that the excessive inflammatory response is an important factor of death in coronavirus disease 2019 (COVID-19) patients. In this study, we propose a deep representation on heterogeneous drug networks, termed DeepR2cov, to discover potential agents for treating the excessive inflammatory response in COVID-19 patients. This work explores the multi-hub characteristic of a heterogeneous drug network integrating eight unique networks. Inspired by the multi-hub characteristic, we design 3 billion special meta paths to train a deep representation model for learning low-dimensional vectors that integrate long-range structure dependency and complex semantic relation among network nodes. Based on the representation vectors and transcriptomics data, we predict 22 drugs that bind to tumor necrosis factor-α or interleukin-6, whose therapeutic associations with the inflammation storm in COVID-19 patients, and molecular binding model are further validated via data from PubMed publications, ongoing clinical trials and a docking program. In addition, the results on five biomedical applications suggest that DeepR2cov significantly outperforms five existing representation approaches. In summary, DeepR2cov is a powerful network representation approach and holds the potential to accelerate treatment of the inflammatory responses in COVID-19 patients. The source code and data can be downloaded from https://github.com/pengsl-lab/DeepR2cov.git.

MOTIVATION: Predicting entity relationship can greatly benefit important biomedical problems. Recently, a large amount of biomedical heterogeneous networks (BioHNs) are generated and offer opportunities for developing network-based learning approaches to predict relationships among entities. However, current researches slightly explored BioHNs-based self-supervised representation learning methods, and are hard to simultaneously capturing local- and global-level association information among entities. RESULTS: In this study, we propose a BioHN-based self-supervised representation learning approach for entity relationship predictions, termed BioERP. A self-supervised meta path detection mechanism is proposed to train a deep Transformer encoder model that can capture the global structure and semantic feature in BioHNs. Meanwhile, a biomedical entity mask learning strategy is designed to reflect local associations of vertices. Finally, the representations from different task models are concatenated to generate two-level representation vectors for predicting relationships among entities. The results on eight datasets show BioERP outperforms 30 state-of-the-art methods. In particular, BioERP reveals great performance with results close to 1 in terms of AUC and AUPR on the drug-target interaction predictions. In summary, BioERP is a promising bio-entity relationship prediction approach. AVAILABILITY AND IMPLEMENTATION: Source code and data can be downloaded from https://github.com/pengsl-lab/BioERP.git. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

Compressive Sensing for Magnetic Resonance Imaging (CS-MRI) aims to reconstruct Magnetic Resonance (MR) images from under-sampled raw data. There are two challenges to improve CS-MRI methods, i.e. designing an under-sampling algorithm to achieve optimal sampling, as well as designing fast and small deep neural networks to obtain reconstructed MR images with superior quality. To improve the reconstruction quality of MR images, we propose a novel deep convolutional neural network architecture for CS-MRI named MRCSNet. The MRCSNet consists of three sub-networks, a compressive sensing sampling sub-network, an initial reconstruction sub-network, and a refined reconstruction sub-network. Experimental results demonstrate that MRCSNet generates high-quality reconstructed MR images at various under-sampling ratios, and also meets the requirements of real-time CS-MRI applications. Compared to state-of-the-art CS-MRI approaches, MRCSNet offers a significant improvement in reconstruction accuracies, such as Peak Signal-to-Noise Ratio (PSNR) and Structural Similarity (SSIM). Besides, it reduces the reconstruction error evaluated by the Normalized Root-Mean-Square Error (NRMSE). The source codes are available at https://github.com/TaihuLight/MRCSNet .

Abstract Before service providers build up an mobile edge computing (MEC) platform, an important issue that needs to be considered is the configuration of computing resources on edge servers. Since the computing resources on an edge server are limited compared with a cloud server and the service provider's deployment budget is limited, it would be unrealistic to equip all edge servers with abundant computing resources. In addition, the edge servers have different computation demands due to their different geographies. Therefore, this article investigates the problem of server configuration optimization in an MEC environment based on a given computation demand statistics of the selected deployment locations. Our strategy is to treat each edge server as an M/M/m queueing model, and then establish the performance and cost models for the system. Two optimization problems, including cost constrained performance optimization, and performance constrained cost optimization are formulated based on our models and solved by a series of fast numerical algorithms. We also conduct extensive numerical simulation examples to show the effectiveness of the proposed algorithms. MEC service providers can use our strategy to get the appropriate type of processor and obtain the optimal processor number for each edge server to achieve two different goals: (1) deliver the highest‐quality services with a given cost constraint; (2) minimize the investment cost with a service‐quality guarantee. Our research is of great significance for service providers to control the tradeoff between investment cost and service quality.

In the environment of cloud computing, the data produced by massive users form a data stream and need to be protected by encryption for maintaining confidentiality. Traditional serial encryption algorithms are poor in performance and consume more energy without considering the property of streams. Therefore, we propose a velocity-aware parallel encryption algorithm with low energy consumption (LECPAES) for streams in cloud computing. The algorithm parallelizes Advanced Encryption Standard (AES) based on heterogeneous many-core architecture, adopts a sliding window to stabilize burst flows, senses the velocity of streams using the thresholds of the window computed by frequency ratios, and dynamically scales the frequency of Graphics Processing Units (GPUs) to lower down energy consumption. The experiments for streams at different velocities and the comparisons with other related algorithms show that the algorithm can reduce energy consumption, but only slightly increases retransmission rate and slightly decreases throughput. Therefore, LECPAES is an excellent algorithm for fast and energy-saving stream encryption.

In recent years, erasure codes have become the de facto standard for data protection in large scale distributed cloud storage systems at the cost of an affordable storage overhead. However, traditional erasure coding schemes, such as Reed-Solomon codes, suffer from high reconstruction cost and I/Os. The recent past has seen a plethora of efforts to optimize the tradeoff between the reconstruction cost, I/Os and storage overhead. Quiet different from all prior studies, in this paper, our erasure coding technique makes the first attempt to take advantage of the unequal failure rates across the disks/nodes to optimize the system reliability and reconstruction performance. Specifically, our proposed technique, the Unequal Failure Protection based Local Reconstruction Code (UFP-LRC) divides the data blocks into several unequal-sized groups with local parities, assigning the data blocks stored on more failure-prone disks/nodes into the smaller-sized group, so as to provide unequal failure protection for each group. In this way, by exploiting the nonuniform local parity degrees, the proposed UFP-LRC enables the data blocks that are stored on more failure-prone disks/nodes to tolerate a greater number of failures while suffering from less repair cost than others, leading to a substantial improvement of the overall reliability and repair performance for cloud storage systems. We perform numerical analysis and build a prototype storage system to verify our approach. The analytical results show that the UFP-LRC technique gradually outperforms LRC along the increase of failure rate ratio. Also, extensive experiments show that, when compared to LRC, UFP-LRC is able to achieve a 10 to 15 percent improvement in throughput, and an 8 to 12 percent reduction in decoding latency, while retaining a comparable overall reliability.

In this paper, we focus on request migration strategies among multi-servers for load balancing. Different from the general load balancing problem, we consider it under a distributed, non-cooperative, and competitive environment. Due to the mentioned characteristics, we view our problem from a game theoretic perspective and formulate it into a non-cooperative game among the multiple servers, in which each server is informed with incomplete information of other servers. For each server, we define its expected response time as a disutility function and try to minimize its value. We also take into account server availability, which impacts the processing capacity of a server and thus its disutility. 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) 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. Finally, we conduct some numerical calculations to verify our theoretical analyses. The experimental results show that our proposed IPA algorithm converges to a Nash equilibrium very quickly and significantly decreases the disutilities of all servers by configuring a proper request migration strategy.

In this paper, we try to design a service mechanism for profit optimizations of both a cloud provider and its multiple users. We consider the problem from a game theoretic perspective and characterize the relationship between the cloud provider and its multiple users as a Stackelberg game, in which the strategies of all users are subject to that of the cloud provider. The cloud provider tries to select and provision appropriate servers and configure a proper request allocation strategy to reduce energy cost while satisfying its cloud users at the same time. We approximate its servers selection space by adding a controlling parameter and configure an optimal request allocation strategy. For each user, we design a utility function which combines the net profit with time efficiency and try to maximize its value under the strategy of the cloud provider. We formulate the competitions among all users as a generalized Nash equilibrium problem (GNEP). We solve the problem by employing variational inequality (VI) theory and prove that there exists a generalized Nash equilibrium solution set for the formulated GNEP. Finally, we propose an iterative algorithm (IA), which characterizes the whole process of our proposed service mechanism. We conduct some numerical calculations to verify our theoretical analyses. The experimental results show that our IA algorithm can benefit both of a cloud provider and its multiple users by configuring proper strategies.

In the parallel computing framework of Hadoop/Spark, data skew is a common problem resulting in performance degradation, such as prolonging of the entire execution time and idle resources. What lies behind this issue is the partition imbalance, that causes significant differences in the amount of data processed by each reduce task. This paper proposes a key reassigning and splitting partition algorithm (SKRSP) to solve the partition skew from the source codes of Spark-core_2.11 project, which considers both the partition balance of the intermediate data and the partition balance after shuffle operators. First, we propose a step-based algorithm for sampling the input data to estimate the general key distribution of entire intermediate data. According to the types of the specific applications, we design two algorithms: hash based key reassigning algorithm (KRHP) and rang based key splitting algorithm (KSRP), which can generate appropriate strategy and implement the skew mitigation in shuffle phase. KKSRP generates the weighted bounds to split intermediate data for the type of sort-based applications while KRHP records these reassigned keys and the new reducers these keys belong to for other applications. Finally, we implement SKRSP in Spark 2.2.0 and evaluate its performance through four benchmarks exhibiting significant data skew: WordCount, Sort, Join, and PageRank. The experimental results verify that our algorithm not only can achieve a better partition balance but also reduce the execution time of reduce tasks effectively.

As a new generation of Public Bicycle-sharing Systems (PBS), the Dockless PBS (DL-PBS) is an important application of cyber-physical systems and intelligent transportation. How to use artificial intelligence to provide efficient bicycle dispatching solutions based on dynamic bicycle rental demand is an essential issue for DL-PBS. In this article, we propose MORL-BD, a dynamic bicycle dispatching algorithm based on multi-objective reinforcement learning to provide the optimal bicycle dispatching solution for DL-PBS. We model the DL-PBS system from the perspective of cyber-physical systems and use deep learning to predict the layout of bicycle parking spots and the dynamic demand of bicycle dispatching. We define the multi-route bicycle dispatching problem as a multi-objective optimization problem by considering the optimization objectives of dispatching costs, dispatch truck's initial load, workload balance among the trucks, and the dynamic balance of bicycle supply and demand. On this basis, the collaborative multi-route bicycle dispatching problem among multiple dispatch trucks is modeled as a multi-agent and multi-objective reinforcement learning model. All dispatch paths between parking spots are defined as state spaces, and the reciprocal of dispatching costs is defined as a reward. Each dispatch truck is equipped with an agent to learn the optimal dispatch path in the dynamic DL-PBS network. We create an elite list to store the Pareto optimal solutions of bicycle dispatch paths found in each action, and finally get the Pareto frontier. Experimental results on the actual DL-PBS show that compared with existing methods, MORL-BD can find a higher quality Pareto frontier with less execution time.

This article is concerned with the multitask multiagent allocation problem via many-objective optimization for multiagent systems (MASs). First, a novel layered MAS model is constructed to address the multitask multiagent allocation problem that includes both the original task simplification and the many-objective allocation. In the first layer of the model, the deep Q-learning method is introduced to simplify the prioritization of the original task set. In the second layer of the model, the modified shift-based density estimation (MSDE) method is put forward to improve the conventional strength Pareto evolutionary algorithm 2 (SPEA2) in order to achieve many-objective optimization on task assignments. Then, an MSDE-SPEA2-based method is proposed to tackle the many-objective optimization problem with objectives including task allocation, makespan, agent satisfaction, resource utilization, task completion, and task waiting time. As compared with the existing allocation methods, the developed method in this article exhibits an outstanding feature that the task assignment and the task scheduling are carried out simultaneously. Finally, extensive experiments are conducted to: 1) verify the validity of the proposed model and the effectiveness of two main algorithms and 2) illustrate the optimal solution for task allocation and efficient strategy for task scheduling under different scenarios.

Recent trend focuses on using heterogeneous graph of things (HGoT) to represent things and their relations in the Internet of Things, thereby facilitating the applying of advanced learning frameworks, i.e., deep learning (DL). Nevertheless, this is a challenging task since the existing DL models are hard to accurately express the complex semantics and attributes for those heterogeneous nodes and links in HGoT. To address this issue, we develop attention-aware encoder-decoder graph neural networks for HGoT, termed as HGAED. Specifically, we utilize the attention-based separate-and-merge method to improve the accuracy, and leverage the encoder-decoder architecture for implementation. In the heart of HGAED, the separate-and-merge processes can be encapsulated into encoding and decoding blocks. Then, blocks are stacked for constructing an encoder-decoder architecture to jointly and hierarchically fuse heterogeneous structures and contents of nodes. Extensive experiments on three real-world datasets demonstrate the superior performance of HGAED over state-of-the-art baselines.

Intelligent ultrasound imaging based on deep learning is one of the important applications in the field of intelligent medical care. In this article, we propose an automatic fetal ultrasound standard plane recognition (FUSPR) model based on deep learning in the Industrial Internet of Things (IIoT) environment. We build a distributed ultrasound data processing and predicting platform by using the IIoT and high-performance computing (HPC) technology. The FUSPR model deployed in the HPC center consists of a convolutional neural network (CNN) component and a recurrent neural network (RNN) component, which learns the spatial and temporal features of the ultrasound video stream by using multitask learning, respectively. The CNN component identifies fetal key anatomical structures from each video frame and accurately recognizes the potential four fetal standard planes. The RNN component obtains the temporal information between adjacent frames, and it realizes precise localization and tracking of fetal organs across frames. In addition, we introduce two feature fusion strategies into the FUSPR model, i.e., CNN fusion and RNN fusion, to fit the spatial sequence and motion representation in the video stream, thereby effectively improving the accuracy and robustness of the model. Extensive experiments conducted on more than 1000 ultrasound videos show that the FUSPR model is superior to the competing baselines in terms of accuracy and performance.

Benefiting from convenient cycling and flexible parking locations, the Dockless Public Bicycle-sharing (DL-PBS) network becomes increasingly popular in many countries. However, redundant and low-utility stations waste public urban space and maintenance costs of DL-PBS vendors. In this article, we propose a Bicycle Station Dynamic Planning (BSDP) system to dynamically provide the optimal bicycle station layout for the DL-PBS network. The BSDP system contains four modules: bicycle drop-off location clustering, bicycle-station graph modeling, bicycle-station location prediction, and bicycle-station layout recommendation. In the bicycle drop-off location clustering module, candidate bicycle stations are clustered from each spatio-temporal subset of the large-scale cycling trajectory records. In the bicycle-station graph modeling module, a weighted digraph model is built based on the clustering results and inferior stations with low station revenue and utility are filtered. Then, graph models across time periods are combined to create a graph sequence model. In the bicycle-station location prediction module, the GGNN model is used to train the graph sequence data and dynamically predict bicycle stations in the next period. In the bicycle-station layout recommendation module, the predicted bicycle stations are fine-tuned according to the government urban management plan, which ensures that the recommended station layout is conducive to city management, vendor revenue, and user convenience. Experiments on actual DL-PBS networks verify the effectiveness, accuracy, and feasibility of the proposed BSDP system.

Multi-task learning plays an important role in face multi-attribute prediction. At present, most researches excavate the shared information between attributes by sharing all convolutional layers. However, it is not appropriate to treat the low-level and high-level features of the face multi-attribute equally, because the high-level features are more biased toward the specific content of the category. In this article, a novel multi-attribute tensor correlation neural network (MTCN) is used to predict face attributes. MTCN shares all attribute features at the low-level layers, and then distinguishes each attribute feature at the high-level layers. To better excavate the correlations among high-level attribute features, each sub-network explores useful information from other networks to enhance its original information. Then a tensor canonical correlation analysis method is used to seek the correlations among the highest-level attributes, which enhances the original information of each attribute. After that, these features are mapped into a highly correlated space through the correlation matrix. Finally, we use sufficient experiments to verify the performance of MTCN on the CelebA and LFWA datasets and our MTCN achieves the best performance compared with the latest multi-attribute recognition algorithms under the same settings.

To accurately detect and track the thyroid nodules in a video is a crucial step in the thyroid screening for identification of benign and malignant nodules in computer-aided diagnosis (CAD) systems. Most existing methods just perform excellent on static frames selected manually from ultrasound videos. However, manual acquisition is labor-intensive work. To make the thyroid screening process in a more natural way with less labor operations, we develop a well-designed framework suitable for practical applications for thyroid nodule detection in ultrasound videos. Particularly, in order to make full use of the characteristics of thyroid videos, we propose a novel post-processing approach, called Cache-Track, which exploits the contextual relation among video frames to propagate the detection results into adjacent frames to refine the detection results. Additionally, our method can not only detect and count thyroid nodules, but also track and monitor surrounding tissues, which can greatly reduce the labor work and achieve computer-aided diagnosis. Experimental results show that our method performs better in balancing accuracy and speed.

CANDECOMP/PARAFAC (CP) decomposition is widely used in various online social network (OSN) applications. However, it is inefficient when dealing with massive and incremental data. Some incremental CP decomposition (ICP) methods have been proposed to improve the efficiency and process evolving data, by updating decomposition results according to the newly added data. The ICP methods are efficient, but inaccurate because of serious error accumulation caused by approximation in the incremental updating. To promote the wide use of ICP, we strive to reduce its cumulative errors while keeping high efficiency. We first differentiate all possible errors in ICP into two types: the cumulative reconstruction error and the prediction error. Next, we formulate two optimization problems for reducing the two errors. Then, we propose several restarting strategies to address the two problems. Finally, we test the effectiveness in three typical dynamic OSN applications. To the best of our knowledge, this is the first work on reducing the cumulative errors of the ICP methods in dynamic OSNs.

Installing a six-dimensional force/torque sensor on an industrial arm for force feedback is a common robotic force control strategy. However, because of the high price of force/torque sensors and the closedness of an industrial robot control system, this method is not convenient for industrial mass production applications. Various types of data generated by industrial robots during the polishing process can be saved, transmitted, and applied, benefiting from the growth of the industrial internet of things (IIoT). Therefore, we propose a constant force control system that combines an industrial robot control system and industrial robot offline programming software for a polishing robot based on IIoT time series data. The system mainly consists of four parts, which can achieve constant force polishing of industrial robots in mass production. (1) Data collection module. Install a six-dimensional force/torque sensor at a manipulator and collect the robot data (current series data, etc.) and sensor data (force/torque series data). (2) Data analysis module. Establish a relationship model based on variant long short-term memory which we propose between current time series data of the polishing manipulator and data of the force sensor. (3) Data prediction module. A large number of sensorless polishing robots of the same type can utilize that model to predict force time series. (4) Trajectory optimization module. The polishing trajectories can be adjusted according to the prediction sequences. The experiments verified that the relational model we proposed has an accurate prediction, small error, and a manipulator taking advantage of this method has a better polishing effect.

Network-wide monitoring is important for many network functions. Due to the need of sampling to reduce high measurement cost, system failure, and unavoidable data transmission loss, network monitoring systems suffer from the incompleteness of network monitoring data. Different from the traditional network monitoring data estimation problem which aims to infer all missing monitoring data entries with incomplete measurement data, we study a challenging problem of inferring the top-$k$largest monitoring data entries. The recent study shows it is promising to more accurately interpolate the missing data with a 3-D tensor compared to that based on a 2-D matrix. Taking full advantage of the multilinear structures, we apply tensor completion to first recover the missing data and then find the top-$k$data entries. To reduce the computational overhead, we propose a novel discrete tensor completion model which uses binary codes to represent the factor matrices. Based on the model, we further propose three novel techniques to speed up the whole top-$k$entry inference process: a discrete optimization algorithm to train the binary factor matrices, bit operations to facilitate quick missing data inference, and simplifying the finding of top-$k$largest entries with binary code partition. In our discrete tensor completion model, only one bit is needed to represent the entry in the factor matrices instead of a real value (32 bits) needed in traditional tensor completion model, thus the storage cost is reduced significantly. To quickly infer the top-$k$largest data entries when measurement data arrive sequentially, we also propose a sliding window based online algorithm using the discrete tensor completion model. Extensive experiments using five real data sets and one synthetic data set demonstrate that compared with the state of art tensor completion algorithms, our discrete tensor completion algorithm can achieve similar top-$k$entry inference accuracy using significantly smaller time and storage space.

Sparse Tucker Decomposition (STD) algorithms learn a core tensor and a group of factor matrices to obtain an optimal low-rank representation feature for the High-Order, High-Dimension, and Sparse Tensor (HOHDST). However, existing STD algorithms face the problem of intermediate variables explosion which results from the fact that the formation of those variables, i.e., matrices Khatri-Rao product, Kronecker product, and matrix-matrix multiplication, follows the whole elements in sparse tensor. The above problems prevent deep fusion of efficient computation and big data platforms. To overcome the bottleneck, a novel stochastic optimization strategy (SGD Tucker) is proposed for STD which can automatically divide the high-dimension intermediate variables into small batches of intermediate matrices. Specifically, SGD Tucker only follows the randomly selected small samples rather than the whole elements, while maintaining the overall accuracy and convergence rate. In practice, SGD Tucker features the two distinct advancements over the state of the art. First, SGD Tucker can prune the communication overhead for the core tensor in distributed settings. Second, the low data-dependence of SGD Tucker enables fine-grained parallelization, which makes SGD Tucker obtaining lower computational overheads with the same accuracy. Experimental results show that SGD Tucker runs at least 2X faster than the state of the art.

Closer to mobile users geographically, mobile edge computing (MEC) can provide some cloud-like capabilities to users more efficiently. This enables it possible for resource-limited mobile users to offload their computation-intensive and latency-sensitive tasks to MEC nodes. For its great benefits, MEC has drawn wide attention and extensive works have been done. However, few of them address task migration problem caused by distributed user mobility, which can't be ignored with quality of service (QoS) consideration. In this article, we study task migration problem and try to minimize the average completion time of tasks under migration energy budget. There are multiple independent users and the movement of each mobile user is memoryless with a sequential decision-making process, thus reinforcement learning algorithm based on Markov chain model is applied with low computation complexity. To further facilitate cooperation among users, we devise a distributed task migration algorithm based on counterfactual multi-agent (COMA) reinforcement learning approach to solve this problem. Extensive experiments are carried out to assess the performance of this distributed task migration algorithm. Compared with no migrating (NM) and single-agent actor-critic (AC) algorithms, the proposed distributed task migration algorithm can achieve up 30-50 percent reduction about average completion time.

With the development of the Internet of Things, the data that needs to be processed is increasing rapidly. Therefore, the collaboration of cloud and edge emerges as the times require. Edge nodes are mainly responsible for collecting data, and decide to process the data locally or offload to cloud data centers. Cloud data centers are suitable for data analysis, model training, and managing edge nodes. In this article, we focus on the task assignment problems in collaborative edge and cloud environments and study it in a distributed, non-cooperative environment. An M/M/1 queueing model is established to characterize the task transmission. Because of the multi-core processors, we set an M/M/C queueing model to characterize the task computation. We consider the problem from the perspective of game theory and formulate it into a non-cooperative game among multi-agents (multiple edge data centers) in which each agent is informed with incomplete information (allocation strategies) of others. For each agent, we define a function of the expected cost of tasks as the disutility function, and minimize it subject to the QoS constraint. We analyze the existence of Nash equilibrium and develop a Greedy Energy-aware Algorithm (GEA) to choose active servers using the Limit Searching Algorithm (LSA) to find the ceiling utilization. Then we propose the Best Response Algorithm (BRA) to optimize the utility function. The convergence of the BRA algorithm has been discussed. Finally, the results demonstrate that the BRA algorithm can get a solution close to Nash equilibrium and reach it quickly.

The Sunway TaihuLight, equipped with 10 million cores, is currently the world's third fastest supercomputer. SpMV is one of core algorithms in many high-performance computing applications. This paper implements a fine-grained design for generic parallel SpMV based on the special Sunway architecture and finds three main performance limitations, i.e., storage limitation, load imbalance, and huge overhead of irregular memory accesses. To address these problems, this paper introduces a customized and accelerative framework for SpMV (CASpMV) on the Sunway. The CASpMV customizes an auto-tuning four-way partition scheme for SpMV based on the proposed statistical model, which describes the sparse matrix structure characteristics, to make it better fit in with the computing architecture and memory hierarchy of the Sunway. Moreover, the CASpMV provides an accelerative method and customized optimizations to avoid irregular memory accesses and further improve its performance on the Sunway. Our CASpMV achieves a performance improvement that ranges from 588.05 to 2118.62 percent over the generic parallel SpMV on a CG (which corresponds to an MPI process) of the Sunway on average and has good scalability on multiple CGs. The performance comparisons of the CASpMV with state-of-the-art methods on the Sunway indicate that the sparsity and irregularity of data structures have less impact on CASpMV.

Large-scale data centers have been widely used for cloud services, and the stability of various cloud services has received additional attention from users. Although service disruptions are not as catastrophic as they once were, their impact might be more extensive than before. These outages may trigger the migration of virtual machines (VMs) located in the failure node. However, the access time of each VM is random, unlike the accident time, which can be predicted. This means that traditional migration caused by service interruptions may result in a large number of unwanted migrations, regardless of the user’s downtime experience. Migration is an expensive process in terms of the resources needed as well as the degradation of application performance during migration. A balance between the recovery time of the service (to minimize the migration resulting from a given placement) and the downtime experience of the users (to minimize the impact of access interruptions) is needed. In this paper, we propose HMGOWM, a hybrid decision-making mechanism for automating the migration of VMs. Our proposed mechanism extends the original VM migration performance cost model, greatly reducing the downtime experience of the users. To achieve high performance and a good load balance, a multi-objective monitoring system for both VMs and physical machine nodes and an adaptive VM migration-scheduling scheme for the OpenStack cloud platform are proposed. Extensive experiment results indicate that the downtime experienced by users can be efficiently reduced and that the implementation of HMGOWM outperforms the original scheduling of the OpenStack cloud platform.

The pricing mechanism of cloud-computing resources is an essential issue for both cloud customers and service providers, especially from the point of multi-provider competition. Although various mechanisms for resource provision are proposed, few studies have focused on multi-attribute resource provision with the objective of improving benefits of both cloud customers and service providers. To address the issue, we propose a price bidding mechanism for multi-attribute cloud-computing resource provision from the perspective of a non-cooperative game, in which the information of each player (customers and providers) is incomplete to others and each player wishes to maximize his/her own benefit. More specifically, considering the fairness pricing competition, we propose a novel and incentive resource provision model referring to the Quality-of-Service (QoS) and the bidding price. Then, combining with the resource provision model, the problem of price bidding is formulated as a game to find a proper price for each cloud provider. We demonstrate the existence of Nash equilibrium solution set for the formulated game model by assuming that the quantity function of provided resources from every provider is continuous. To find a Nash equilibrium solution, we propose an Equilibrium Solution Iterative (ESI) algorithm, which is proved to converge to a Nash equilibrium. Finally, a Near-equalization Price Bidding (NPB) algorithm is proposed to modify the obtained Nash equilibrium solution. Extensive simulated experiments results and the comparison experiments with the state-of-the-art and benchmark solutions validate and show the feasibility of the proposed method.

End-to-end latency is sensitive for user-interactive neural network (NN) services on clouds. For periods of high request load, co-locating multiple NN requests has the potential to reduce end-to-end latency. However, current batch-based accelerators lack request-level parallelism support, leaving the queuing time non-optimized. Meanwhile, naively partitioning resources for simultaneous requests suffers from longer execution time as well as lower resource efficiency because different applications utilize separate resources without sharing. To effectively reduce the end-to-end latency for real-time NN requests, we propose CoExe architecture, equipped with a pipeline implementation of a sparsity-driven real-time co-execution model. By leveraging the non-trivial amount of sparse operations during concurrent NNs execution, the end-to-end latency is decreased by up to 12.3× and 2.4× over Eyeriss-like and SCNN at peak workload mode. Besides, we propose row cross (RC) dataflow to reduce data movement cost, and avoid memory duplication.

Clustering gene expression profiles is a fundamental task in the genome and biomedical research. With the development of RNA-seq and gene chip technology, mass gene expression profile data has been generated, which puts forward two requirements for related research of gene expression profile: i) accurate analysis of drug R&D requires high accuracy of similarity analysis, ii) large-scale analysis of data requires as little running time as possible. We propose a faster, more accurate method called DeepCDNet, which is based on the framework of the Siamese network. DeepCDNet uses the DenseNet structure and optimized loss function to achieve rapid convergence, and the similarity between expression spectra is calculated by a cosine function. The experiment results show that: i) our method breaks through the limitation of high dimensions of gene expression profile and can quickly and accurately learn the required gene characteristics, ii) The accuracy of our method in similarity analysis is greatly improved, iii) as the dimension of data increases, the advantage of our method on time cost gradually becomes more prominent, and time consumption is less.

Path explosion and constraint solving are two challenges to symbolic execution's scalability. Symbolic execution explores the program's path space with a searching strategy and invokes the underlying constraint solver in a black-box manner to check the feasibility of a path. Inside the constraint solver, another searching procedure is employed to prove or disprove the feasibility. Hence, there exists the problem of double searchings in symbolic execution. In this paper, we propose to unify the double searching procedures to improve the scalability of symbolic execution. We propose Multiplex Symbolic Execution (MuSE) that utilizes the intermediate assignments during the constraint solving procedure to generate new program inputs. MuSE maps the constraint solving procedure to the path exploration in symbolic execution and explores multiple paths in one time of solving. We have implemented MuSE on two symbolic execution tools (based on KLEE and JPF) and three commonly used constraint solving algorithms. The results of the extensive experiments on real-world benchmarks indicate that MuSE has orders of magnitude speedup to achieve the same coverage.

Answer selection plays a crucial role in natural language processing. and thus has received much attention. Many recent works treat it as an ad-hoc retrieval problem where ranking optimization accounts for a large proportion. Previous works mainly consider the similarity between answer and question, but rarely utilize similarity and dissimilarity relationship in the answers candidate set. In this paper, we propose a similarity aggregation method to rerank the results produced by different baseline neural networks. The key idea of similarity aggregation is that true matches should not only similar to other true matches, but also dissimilar with false matches, and inspired by multi-view verification, the true answers should have the same ranking to the question in different baseline methods and false answers are the same. The empirical results, from the public benchmark task of answer selection, demonstrate that our method has significant improvement over the baseline methods.

Serendipity recommendation has attracted more and more attention in recent years; it is committed to providing recommendations which could not only cater to users’ demands but also broaden their horizons. However, existing approaches usually measure user-item relevance with a scalar instead of a vector, ignoring user preference direction, which increases the risk of unrelated recommendations. In addition, reasonable explanations increase users’ trust and acceptance, but there is no work to provide explanations for serendipitous recommendations. To address these limitations, we propose a Directional and Explainable Serendipity Recommendation method named DESR. Specifically, we extract users’ long-term preferences with an unsupervised method based on GMM (Gaussian Mixture Model) and capture their short-term demands with the capsule network at first. Then, we propose the serendipity vector to combine long-term preferences with short-term demands and generate directionally serendipitous recommendations with it. Finally, a back-routing scheme is exploited to offer explanations. Extensive experiments on real-world datasets show that DESR could effectively improve the serendipity and explainability, and give impetus to the diversity, compared with existing serendipity-based methods.

BACKGROUND: Biomedical named entity recognition (BioNER) is an important task for understanding biomedical texts, which can be challenging due to the lack of large-scale labeled training data and domain knowledge. To address the challenge, in addition to using powerful encoders (e.g., biLSTM and BioBERT), one possible method is to leverage extra knowledge that is easy to obtain. Previous studies have shown that auto-processed syntactic information can be a useful resource to improve model performance, but their approaches are limited to directly concatenating the embeddings of syntactic information to the input word embeddings. Therefore, such syntactic information is leveraged in an inflexible way, where inaccurate one may hurt model performance. RESULTS: In this paper, we propose BIOKMNER, a BioNER model for biomedical texts with key-value memory networks (KVMN) to incorporate auto-processed syntactic information. We evaluate BIOKMNER on six English biomedical datasets, where our method with KVMN outperforms the strong baseline method, namely, BioBERT, from the previous study on all datasets. Specifically, the F1 scores of our best performing model are 85.29% on BC2GM, 77.83% on JNLPBA, 94.22% on BC5CDR-chemical, 90.08% on NCBI-disease, 89.24% on LINNAEUS, and 76.33% on Species-800, where state-of-the-art performance is obtained on four of them (i.e., BC2GM, BC5CDR-chemical, NCBI-disease, and Species-800). CONCLUSION: The experimental results on six English benchmark datasets demonstrate that auto-processed syntactic information can be a useful resource for BioNER and our method with KVMN can appropriately leverage such information to improve model performance.