With the widespread deployment of cross-modal retrieval in real-world scenarios, ensuring robustness against adversarial attacks is increasingly critical. Remarkably, deep cross-modal hashing is highly vulnerable to adversarial attacks due to its discrete nature and low-dimensional hash codes, while existing defense methods often fail to suppress perturbations embedded in vulnerable features and lack the capacity to model modality-specific structural differences, resulting in suboptimal adversarial robustness. To address these challenges, we propose a novel Disentangled Representation-Focused Generative Defense (DRFGD) framework for attack-tolerant cross-modal hashing. Without altering the structure of retrieval model, DRFGD defends against adversarial attacks by disentangling input representations into adversarial-robust and adversarial-vulnerable components, by an efficient dual-branch semantic-aware encoder. Guided by such disentangled robust features, an attack-tolerant generative module is seamlessly designed to synthesize semantically aligned and perturbation-resilient examples for robust adversarial training, thereby significantly promoting collaborative defense robustness to attackers. Consequently, the semantically consistent hash codes can be well obtained to enhance adversarial robustness in complex cross-modal attacking scenarios. Extensive experiments on public benchmarks demonstrate that DRFGD substantially improves retrieval robustness under various attacking scenarios, and shows its improved defense performance in comparison with the SOTA works.

Existing adaptive learning systems struggle to simultaneously achieve deep personalization, dynamic adaptability, and content trustworthiness, particularly in logically rigorous STEM fields where Large Language Models (LLMs) are prone to "hallucination".This paper introduces LEARNERCOMPASS (Cognitive Multimodel Planning & Adaptive System), an integrated, end-to-end framework for adaptive learning.At its core, the framework features a novel multi-model path planning algorithm that orchestrates and fuses the outputs of heterogeneous LLM experts to generate and optimize learning sequences.To enable deep personalization, we design a dynamic cognitive diagnosis module that employs an innovative encoderdecoder architecture to generate precise, multidimensional cognitive state vectors for learners.To ensure trustworthiness, the system leverages an adaptively constructed dynamic knowledge graph and a Graph-RAG mechanism to provide factual anchors and logical constraints for LLM reasoning, thereby mitigating hallucinations.Extensive experiments demonstrate that LEARNERCOMPASS significantly outperforms state-of-the-art baselines in generating high-quality personalized learning paths.Furthermore, ablation studies validate the critical contributions of our dynamic cognitive diagnosis and multi-model planning components.

Xueyang Zhou, Weidong Wang, Lin Lu, Jiawen Shi, Guiyao Tie, Xu Yongtian, Lixing Chen, Pan Zhou, Neil Zhenqiang Gong, Lichao Sun. Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). 2026.

The integration of Domain Knowledge Graphs (DKG) into Retrieval-Augmented Generation has emerged as a promising approach for constructing Specialized Large Language Models (spLLMs).On account of the scarcity of high-quality DKGs, existing approaches employ an evolutionary framework, wherein the DKG is continuously evolved alongside its utilization for enhancing the LLM. Yet, these methods face two key limitations: 1) heavy reliance on costly expert knowledge, and 2) neglect of the connection between the LLM's inherent knowledge and external expert knowledge. To address these issues, this paper introduces Epistemic Cognition-enhanced Specialized LLMs (EC-spLLM), a novel evolutionary framework that instills epistemic cognition into the LLM to systematically exploit both its internal knowledge and expert knowledge. At the core of EC-spLLM lies the Hallucination Detection-based Epistemic Cognition (HDEC) mechanism, which assesses the reliability of LLM-generated responses using the LLM's self cognition and hallucination detection. This assessment ability enables EC-spLLM to selectively adopt either the expert-provided golden answer or a reliable LLM-generated answer during DKG evolution, thereby reducing dependence on experts and bridging internal and external knowledge sources to enhance performance. We conducted extensive experiments on five datasets spanning five domain, e.g., emotional sociology, biology, ect. Results show that EC-spLLM reduces the usage of golden answers by an average of 67% while retaining 97.2% of the accuracy achieved by the SOTA method, and outperforms all other baselines.

Cross-view geo-localization (CVGL) establishes correspondences between ground-level and satellite images of the same geographic location, serving as a fundamental technology for smart city applications, including autonomous navigation, urban planning, and location-based services. Current CVGL approaches fall into two categories: feature-based methods achieve superior performance through 2D representation learning but lack interpretability. Spatial-based methods provide geometric understanding and interpretable matching but suffer from limited spatial modeling and weak cross-view alignment, leading to lower performance. We reformulate CVGL from a spatial perspective and propose an auxiliary task-enhanced network. The network captures spatial semantics and provides explicit alignment processes with visualizable results. We introduce an auxiliary spatial semantic alignment (SSA) task that learns spatial structure via vision foundation models (VFM) and BEV transformation to enhance the primary CVGL task. The primary task captures visual semantics, including texture and appearance. Within this unified framework with shared encoders, the primary task enriches the learned embeddings by fusing spatial structure with visual semantics, yielding spatially complete representations. Extensive experiments on three standard CVGL benchmarks demonstrate that our method significantly surpasses previous spatial-based approaches while maintaining competitive performance with state-of-the-art (SOTA) feature-based methods, achieving 98.48% R@1 on CVUSA and 71.05% R@1 on CVACT\_test. We provide comprehensive analyses through pixel-level activation maps and feature-space UMAP visualizations to validate both effectiveness and interpretability.

Identifying intrusions from massive logs has long been a great challenge. To address this issue, this paper proposes Pontus, a novel host-based intrusion detection method via accurate provenance clustering, efficient graph serialization and classification with minimum provenance loss. Pontus first utilizes a novel multi-round label propagation algorithm (MLPA) based on overlapping community discovery to cluster the behavior instances that constitute user behavior accurately. In this way, Pontus can analyze behavior instances to extract behavior features effectively while reducing the analysis workload. Then, Pontus enables efficient graph serialization via neighbor node aggregation to convert the behavior instance into vectors while maximizing the retention of provenance information. Finally, Pontus uses a hybrid method that combines the convolutional autoencoder with Bisecting Kmeans clustering to accurately extract the provenance features of behavior instances to identify host-based intrusions. The experimental results show that compared with the state-of-the-art methods, Pontus’s accuracy increases by an average of 0.243, and F1-score increases by an average of 0.201, with small runtime overheads.

With the maturity of depth sensors in various 3D safety-critical applications, 3D point cloud models have been shown to be vulnerable to adversarial attacks. Almost all existing 3D attackers simply follow the white-box or black-box setting to iteratively update coordinate perturbations based on back-propagated or estimated gradients. However, these methods are hard to deploy in real-world scenarios (no model details are provided) as they severely rely on parameters or output logits of victim models. To this end, we propose point cloud attacks from a more practical setting,i.e., hard-label black-box attack, in which attackers can only access the prediction label of 3D input. We introduce a novel 3D attack method based on a new spectrum-aware decision boundary algorithm to generate high-quality adversarial samples. In particular, we first construct a class-aware model decision boundary, by developing a learnable spectrum-fusion strategy to adaptively fuse point clouds of different classes in the spectral domain, aiming to craft their intermediate samples without distorting the original geometry. Then, we devise an iterative coordinate-spectrum optimization method with curvature-aware boundary search to move the intermediate sample along the decision boundary for generating adversarial point clouds with trivial perturbations. Experiments demonstrate that our attack competitively outperforms existing white/black-box attackers in terms of attack performance and adversary quality.

With the growth of cloud computing infrastructure, its cost-efficient paradigm is driving a growing wave of small and medium-sized enterprises to migrate data and services to cloud based platforms. However, due to privacy concerns, data encryption prior to outsourcing is an important means of protection, which in turn requires performing queries over the encrypted data. While several approaches do offer support for privacy preserving skyline or top-k queries, they typically struggle with efficiency when extended to secure top-k dominating queries, due to the inherent nature of combining the advantages of both top k and skyline queries. Nevertheless, this nature renders them as a more practical and promising alternative for location-based services. To address this, we introduce STLD, a secure top-k location-based dominating query scheme. Specifically, we develop an innovative index structure called Secure Aggregate R-tree (SAR-tree) by utilizing the Paillier cryptosystem and introducing meticulously crafted noise, while also incorporating principles from aggregate R-trees and semi-blind R-trees. Leveraging this structure, we propose a series of secure sub-protocols to facilitate top-k dominating queries, accompanied by optimization techniques to mitigate latency associated with computationally intensive dominating operations. Given an encrypted query, STLD not only efficiently answers the query but also guarantees the privacy of data(sets), results, queries and access patterns. Finally, STLD undergoes rigorous theoretical security and complexity analysis, complemented by empirical evaluations that demonstrate its performance and feasibility, achieving a reduction in query cost by 40%-60% compared to multiple competing methods.

Large vision-language models (LVLMs) have demonstrated remarkable capabilities across a wide range of multimodal understanding and reasoning tasks. However, recent research shows that LVLMs are susceptible to adversarial examples. Existing attackers either optimize the perturbations on the visual input or manipulate prompts to fool the LVLM models, requiring extensive design and engineering on these adversarial manipulations. While straightforward visual transformation can boast training generalization-ability, its potential risks to LVLMs in terms of safety and trustworthiness have been largely neglected. In this paper, we ask an intriguing question:can simple yet easy-to-implement adversarial visual transformations be utilized to attack the LVLM models?Motivated by this research gap and new attack setting, we propose the first comprehensive assessment of LVLMs’ adversarial robustness to visual transformations by testing LVLMs’ resilience to all possible transformation operations. Our empirical observations suggest that with the appropriate combination of the most harmful transformations, we can build transformation-based attacks more adversarial to the LVLM models. Moreover, adversarial learning of visual transformations is further introduced to adaptively apply the malicious impacts of all potentially harmful transformations to the raw images via gradient approximation for improving the attack effectiveness and imperceptibility. We hope that this study can provide deeper insights into the potential vulnerability of LVLMs to adversarial visual transformations.

Betweenness centrality is one of the key centrality measures in many applications including community detections in biological networks, vulnerability detections in communication networks, misinformation filtering in social networks, etc. The top-K group betweenness centrality problem is to find a group of K nodes from a network so that the total fraction of shortest paths that pass through the K nodes is maximized. Existing studies proposed randomized sampling algorithms for the problem. We notice that the existing studies ensured that, the maximum deviation of the estimated centrality of every group from its expectation is no greater than a small given threshold for all potential groups with no more than K nodes, thereby generating too many samples, as the number of such groups is prohibitively large. In contrast, in this paper we first devise a novel algorithm that enables to estimate the centrality of a tentative group adaptively, and the algorithm immediately stops once the centrality is large enough; otherwise, the algorithm uses more samples to find a better group. We then theoretically show that, even the proposed algorithm uses much less samples, it still can find a performance-guaranteed group with high probability. Experimental results with real-world networks demonstrate that the number of samples used by the proposed algorithm is up to 36 times smaller than the state-of-the-art, while the centrality of the group found by the algorithm is no more than 4.5% smaller than the latter.

Web 3.0-enabled edge computing provides a promising foundation for decentralized application (DApp) deployment. However, optimizing execution modes for interdependent task patterns remains challenging due to high computational overhead and blockchain consensus latency. This paper proposes an elastic hybrid architecture for Web 3.0-Enabled Edge Computing Systems (EDGEWEB3.0) that integrates on-chain and off-chain execution. DApp contracts are modeled as Directed Acyclic Graphs (DAGs) and partitioned into dependent task patterns, enabling a structured decomposition of execution. We formulate a DApp task scheduling problem and introduce the Decision Transformer-based Dependent Scheduling (DTDS) algorithm. DTDS employs masked self-attention to capture DAG dependencies and incorporates regularization terms to enforce constraints on service delays, gas costs, and computing capacity. We implement a real-world Web 3.0 testbed based on Ethereum, Goerli, and zkSync. Experimental results demonstrate that DTDS consistently outperforms state-of-the-art baselines in gas costs and service delays, providing a scalable solution for DApp execution in environments.

Bandit with infinitely many arms (i.e.,X-armed bandit) is an important variant of multi-armed stochastic bandits, which is useful to model different networking problems under both wired and wireless settings, e.g., online caching, dynamic channel/power allocation, rate adaption. However, the problem becomes challenging when the characteristic of the networking setting is affected by the side information (i.e., context) and distributed behavior. In this paper, we identify and study a novel problem, decentralized contextualX-armed bandit, whereNagents collaboratively solve the problem within time spanT. The problem is nontrivial because the infinite arms challenge and statistical information consensus issue make our setting go beyond a simple combination ofX-armed bandits and multi-agent bandits. We develop a decentralized arm selection algorithm, called MACXUCB, by elaborating the contextual covering tree technique with a novelgossip communication protocol, which allows each agent to communicate efficiently with her neighbors. We prove that MACXUCB achieves a sublinear regret upper bound Õ(D1/2dX+2dY+4NdX+dY+1.5/dX+dY+2TdX+dY+1/dX+dY+2) given aggregation periodD, and the covering dimensionsdXanddYof arm and context spaces, which asymptomatically matches the lower bound Ω((DN)1/dX+dY+2TdX+dY+1/dX+dY+2) up to a time-dependent factor. Moreover, MACXUCB enjoys a sublinear communication complexity Õ(NDT0.5/(1–p)) when tuning parameterp∈ (0, 1/2). Finally, we carry out experiments to verify the performance of our MACXUCB. The results show the effectiveness and efficiency of our MACXUCB.

In the evolving landscape of Web 3.0, deploying and scheduling decentralized applications (DApps) presents significant challenges due to the complexity of heterogeneous nodes, edges, and their intricate interactions. Traditional approaches, particularly graph-based reinforcement learning (RL) methods, often rely on homogeneous graphs, which fail to capture the diverse relationships inherent in heterogeneous Web 3.0 environments. This limitation results in inefficient resource allocation, suboptimal task scheduling, and unclear security requirements. To address these issues, this paper introduces the Heterogeneous Graph Deployment Scheduler (HGDS), a novel framework that leverages Heterogeneous Graph Neural Networks (HGNNs) to model users, edge servers, and DApp tasks within Web 3.0 environments, and incorporates RL to optimize DApp scheduling policies. HGDS captures heterogeneity in Web 3.0 environments by jointly modeling node–edge interactions and heterogeneous edge relationships, enabling the generation of dynamic, task-aware embeddings that integrate both node and edge features. RL is further employed to adaptively optimize scheduling and resource allocation based on real-time network feedback. Experiments on a Web 3.0 testbed show that HGDS outperforms baseline methods by 12.2% in reward, while reducing service delay and gas consumption.

Recent years have witnessed significant progress in developing deep learning-based models for automated code completion. Examples of such models include CodeGPT and StarCoder. These models are typically trained from a large amount of source code collected from open source communities such as GitHub. Although using source code in GitHub has been a common practice for training deep-learning-based models for code completion, it may induce some legal and ethical issues such as copyright infringement. In this article, we investigate the legal and ethical issues of current neural code completion models by answering the following question: Is my code used to train your neural code completion model ? To this end, we tailor a membership inference approach (termed CodeMI ) that was originally crafted for classification tasks to a more challenging task of code completion. In particular, since the target code completion models perform as opaque black boxes, preventing access to their training data and parameters, we opt to train multiple shadow models to mimic their behavior. The acquired posteriors from these shadow models are subsequently employed to train a membership classifier. After that, the membership classifier can be effectively employed to deduce the membership status of a given code sample based on the output of a target code completion model. We comprehensively evaluate the effectiveness of this adapted approach across a diverse array of neural code completion models (i.e., LSTM-based, CodeGPT, CodeGen, and StarCoder). Experimental results demonstrate that our approach effectively detects data membership, achieving accuracies of 0.842 and 0.730 for LSTM-based and CodeGPT models, respectively. Interestingly, our experiments also show that the data membership of current large language models of code, e.g., CodeGen and StarCoder, is difficult to detect, leaving ample space for further improvement. Finally, we also try to explain the findings from the perspective of model memorization.

Digital twin is emerging as a key technology to monitor the status of complex industry systems. Valuable insights, such as running statuses and anomalies, can be analyzed from the collected system status timely. Considering that the data updating from each system component (known as a physical object) to its digital twin is performed continuously, timely and accurate streaming analytics based on machine learning models is a key technology to analyze such data efficiently. In this paper, we focus on enabling low-delay yet highly-accurate streaming analytics for digital twin applications in mobile edge computing (MEC) networks. Specifically, we formulate a fundamental optimization problem of digital twin placements and model selections for streaming analytics, with the aim of minimizing both the analytic loss and the processing delay. To this end, we first consider the problem with a single query, for which, we propose an approximation algorithm with provable approximation ratio for a special case, and then devise an efficient algorithm for the original problem with a single query. We then study the online digital twin placement and model selection problem for streaming analytics with multiple queries under real scenarios, where resource demands of arrival queries and resource availability of MEC network are uncertain. We propose an online learning algorithm with a bounded regret to make admission policies. We finally evaluate the performance of the proposed algorithms by extensive simulations. Results show that the weighted sums of the total processing delay and the cumulative loss in the solution delivered by the proposed algorithms outperform their counterparts by 12.5% with a single query and 13.3% with multiple queries, respectively.

Data stream processing is a functionality of various AI applications to obtain continuous insights from data streams. Serverless edge computing (SEC) is a key solution for implementing data stream processing requests by deploying serverless functions into cloudlets. However, existing data stream processing methods focus more on processing delay, ignoring fault tolerance and complex dependencies among functions, resulting in critical events being missed in the event of any fault and processing inefficiency. Besides, due to the uncertainty of data streams, existing function deployment methods may not be suitable for their newly changed data rates, causing resource waste or shortages. To address these problems, we first propose an optimization framework to enable efficient and fault tolerant function deployment, such that the delay of data stream processing is minimized while meeting its fault tolerant requirements and resource capacity constraints of cloudlets in an SEC network. We then design an online learning algorithm that predicts data rate changes through a multi-timescale machine learning method and proactively adjusts instance locations and numbers to absorb data rate uncertainty. Experimental results in a real test-bed show that our proposed algorithms outperform their counterparts by 13.5% on the average delay and 26.3% on the average fault tolerance.

Recent studies have revealed the vulnerability of graph neural networks (GNNs) to adversarial poisoning attacks on node classification tasks. Current defensive methods require substituting the original GNNs with defense models, regardless of the original's type. This approach, while targeting adversarial robustness, compromises the enhancements developed in prior research to boost GNNs' practical performance. Here we introduce Grimm, the first plug-and-play defense model. With just a minimal interface requirement for extracting features from any layer of the protected GNNs, Grimm is thus enabled to seamlessly rectify perturbations. Specifically, we utilize the feature trajectories (FTs) generated by GNNs, as they evolve through epochs, to reflect the training status of the networks. We then theoretically prove that the FTs of victim nodes will inevitably exhibit discriminable anomalies. Consequently, inspired by the natural parallelism between the biological nervous and immune systems, we construct Grimm, a comprehensive artificial immune system for GNNs. Grimm not only detects abnormal FTs and rectifies adversarial edges during training but also operates efficiently in parallel, thereby mirroring the concurrent functionalities of its biological counterparts. We experimentally confirm that Grimm offers four empirically validated advantages: 1) Harmlessness, as it does not actively interfere with GNN training; 2) Parallelism, ensuring monitoring, detection, and rectification functions operate independently of the GNN training process; 3) Generalizability, demonstrating compatibility with mainstream GNNs such as GCN, GAT, and GraphSAGE; and 4) Transferability, as the detectors for abnormal FTs can be efficiently transferred across different systems for one-step rectification.

Given some video-query pairs with untrimmed videos and sentence queries, temporal sentence grounding (TSG) aims to locate query-relevant segments in these videos. Although previous respectable TSG methods have achieved remarkable success, they train each video-query pair separately and ignore the relationship between different pairs. To this end, in this paper, we pose a brand-new setting: Multi-Pair TSG, which aims to co-train these pairs. We propose a novel video-query co-training approach, Multi-Thread Knowledge Transfer Network, to locate a variety of video-query pairs effectively and efficiently. Firstly, we mine the spatial and temporal semantics across different queries to cooperate with each other. To learn intra- and inter-modal representations simultaneously, we design a cross-modal contrast module to explore the semantic consistency by a self-supervised strategy. To fully align visual and textual representations between different pairs, we design a prototype alignment strategy to 1) match object prototypes and phrase prototypes for spatial alignment, and 2) align activity prototypes and sentence prototypes for temporal alignment. Finally, we develop an adaptive negative selection module to adaptively generate a threshold for cross-modal matching. Extensive experiments show the effectiveness and efficiency of our proposed method.

End-to-end training with global optimization have popularized graph neural networks (GNNs) for node classification, yet inadvertently introduced vulnerabilities to adversarial edge-perturbing attacks. Adversaries can exploit the inherent opened interfaces of GNNs' input and output, perturbing critical edges and thus manipulating the classification results. Current defenses, due to their persistent utilization of global-optimization-based end-to-end training schemes, inherently encapsulate the vulnerabilities of GNNs. This is specifically evidenced in their inability to defend against targeted secondary attacks. In this paper, we propose the Graph Agent Network (GAgN) to address the aforementioned vulnerabilities of GNNs. GAgN is a graph-structured agent network in which each node is designed as an 1-hop-view agent. Through the decentralized interactions between agents, they can learn to infer global perceptions to perform tasks including inferring embeddings, degrees and neighbor relationships for given nodes. This empowers nodes to filtering adversarial edges while carrying out classification tasks. Furthermore, agents' limited view prevents malicious messages from propagating globally in GAgN, thereby resisting global-optimization-based secondary attacks. We prove that single-hidden-layer multilayer perceptrons (MLPs) are theoretically sufficient to achieve these functionalities. Experimental results show that GAgN effectively implements all its intended capabilities and, compared to state-of-the-art defenses, achieves optimal classification accuracy on the perturbed datasets.

Decentralized training has become a resourceefficient framework to democratize the training of large language models (LLMs).However, the privacy risks associated with this framework, particularly due to the potential inclusion of sensitive data in training datasets, remain unexplored.This paper identifies a novel and realistic attack surface: the privacy leakage from training data in decentralized training, and proposes activation inversion attack (AIA) for the first time.AIA first constructs a shadow dataset comprising text labels and corresponding activations using public datasets.Leveraging this dataset, an attack model can be trained to reconstruct the training data from activations in victim decentralized training.We conduct extensive experiments on various LLMs and publicly available datasets to demonstrate the susceptibility of decentralized training to AIA.These findings highlight the urgent need to enhance security measures in decentralized training to mitigate privacy risks in training LLMs.

Model merging for Large Language Models (LLMs) directly fuses the parameters of different models finetuned on various tasks, creating a unified model for multi-domain tasks.However, due to potential vulnerabilities in models available on open-source platforms, model merging is susceptible to backdoor attacks.In this paper, we propose Merge Hijacking, the first backdoor attack targeting model merging in LLMs.The attacker constructs a malicious upload model and releases it.Once a victim user merges it with any other models, the resulting merged model inherits the backdoor while maintaining utility across tasks.Merge Hijacking defines two main objectives-effectiveness and utility-and achieves them through four steps.Extensive experiments demonstrate the effectiveness of our attack across different models, merging algorithms, and tasks.Additionally, we show that the attack remains effective even when merging real-world models.Moreover, our attack demonstrates robustness against two inference-time defenses (Paraphrasing and CLEANGEN) and one training-time defense (Fine-pruning).

Multi-modal large language models (MLLMs) extend large language models (LLMs) to process multi-modal information, enabling them to generate responses to image-text inputs. MLLMs have been incorporated into diverse multimodal applications, such as autonomous driving and medical diagnosis, via plug-and-play without fine-tuning. This deployment paradigm increases the vulnerability of MLLMs to backdoor attacks. However, existing backdoor attacks against MLLMs achieve limited effectiveness and stealthiness. In this work, we propose BadToken, the first token-level backdoor attack to MLLMs. BadToken introduces two novel backdoor behaviors: Token-substitution and Token-addition, which enable flexible and stealthy attacks by making token-level modifications to the original output for backdoored inputs. We formulate a general optimization problem that considers the two backdoor behaviors to maximize the attack effectiveness. We evaluate BadToken on two open-source MLLMs and various tasks. Our results show that our attack maintains the model’s utility while achieving high attack success rates and stealthiness. We also show the real-world threats of BadToken in two scenarios, i.e., autonomous driving and medical diagnosis. Furthermore, we consider defenses including fine-tuning and input purification. Our results highlight the threat of our attack.

Model merging has emerged as a promising approach for updating large language models (LLMs) by integrating multiple domain-specific models into a cross-domain merged model. Despite its utility and plug-and-play nature, unmonitored mergers can introduce significant security vulnerabilities, such as backdoor attacks and model merging abuse. In this paper, we identify a novel and more realistic attack surface where a malicious merger can extract targeted personally identifiable information (PII) from an aligned model with model merging. Specifically, we propose Merger-as-a-Stealer, a two-stage framework to achieve this attack: First, the attacker fine-tunes a malicious model to force it to respond to any PII-related queries. The attacker then uploads this malicious model to the model merging conductor and obtains the merged model. Second, the attacker inputs direct PII-related queries to the merged model to extract targeted PII. Extensive experiments demonstrate that Merger-as-a-Stealer successfully executes attacks against various LLMs and model merging methods across diverse settings, highlighting the effectiveness of the proposed framework. Given that this attack enables character-level extraction for targeted PII without requiring any additional knowledge from the attacker, we stress the necessity for improved model alignment and more robust defense mechanisms to mitigate such threats.

Intrusion Detection Systems (IDS) play a crucial role in network security defense.However, a significant challenge for IDS in training detection models is the shortage of adequately labeled malicious samples.To address these issues, this paper introduces a novel semi-supervised framework GANGRL-LLM, which integrates Generative Adversarial Networks (GANs) with Large Language Models (LLMs) to enhance malicious code generation and SQL Injection (SQLi) detection capabilities in few-sample learning scenarios.Specifically, our framework adopts a collaborative training paradigm where: (1) the GAN-based discriminator improves malicious pattern recognition through adversarial learning with generated samples and limited real samples; and (2) the LLM-based generator refines the quality of malicious code synthesis using reward signals from the discriminator.The experimental results demonstrate that even with a limited number of labeled samples, our training framework is highly effective in enhancing both malicious code generation and detection capabilities.This dual enhancement capability offers a promising solution for developing adaptive defense systems capable of countering evolving cyber threats.

In spite of achieving significant progress in recent years, Large Vision-Language Models (LVLMs) are proven to be vulnerable to adversarial examples. Therefore, there is an urgent need for an effective adversarial attack to identify the deficiencies of LVLMs in security-sensitive applications. However, existing LVLM attackers generally optimize adversarial samples against a specific textual prompt with a certain LVLM model, tending to overfit the target prompt/network and hardly remain malicious once they are transferred to attack a different prompt/model. To this end, in this paper, we propose a novel Imperceptible Transfer Attack (ITA) against LVLMs to generate prompt/model-agnostic adversarial samples to enhance such adversarial transferability while further improving the imperceptibility. Specifically, we learn to apply appropriate visual transformations on image inputs to create diverse input patterns by selecting the optimal combination of operations from a pool of candidates, consequently improving adversarial transferability. We conceptualize the selection of optimal transformation combinations as an adversarial learning problem and employ a gradient approximation strategy with noise budget constraints to effectively generate imperceptible transferable samples. Extensive experiments on three LVLM models and two widely used datasets with three tasks demonstrate the superior performance of our ITA.

Betweenness centrality is one of the key centrality measures in many applications including community detections in biological networks, vulnerability detections in communication networks, misinformation filtering in social networks, etc. The top-$K$group betweenness centrality problem is to find a group of$K$nodes from a network so that the total fraction of shortest paths that pass through the$K$nodes is maximized. Existing studies proposed randomized sampling algorithms for the problem. We notice that the existing studies ensured that, the maximum deviation of the estimated centrality of every group from its expectation is no greater than a small given threshold for all potential groups with no more than$K$nodes, thereby generating too many samples, as the number of such groups is prohibitively large. In contrast, in this paper we first devise a novel algorithm that enables to estimate the centrality of a tentative group adaptively, and the algorithm immediately stops once the centrality is large enough; otherwise, the algorithm uses more samples to find a better group. We then theoretically show that, even the algorithm uses much less samples, it still can find a performance-guaranteed group with a large success probability. Experimental results with real-world networks demonstrate that the number of samples used by the proposed algorithm is from 2 to 18 times smaller than the state-of-the-art, while the centrality of the group found by the algorithm is no more than 4% smaller than the latter.

Many real-world user queries (e.g. *"How do to make egg fried rice?"*) could benefit from systems capable of generating responses with both textual steps with accompanying images, similar to a cookbook. Models designed to generate interleaved text and images face challenges in ensuring consistency within and across these modalities. To address these challenges, we present ISG, a comprehensive evaluation framework for interleaved text-and-image generation. ISG leverages a scene graph structure to capture relationships between text and image blocks, evaluating responses on four levels of granularity: holistic, structural, block-level, and image-specific. This multi-tiered evaluation allows for a nuanced assessment of consistency, coherence, and accuracy, and provides interpretable question-answer feedback. In conjunction with ISG, we introduce a benchmark, ISG-Bench, encompassing 1,150 samples across 8 categories and 21 subcategories. This benchmark dataset includes complex language-vision dependencies and golden answers to evaluate models effectively on vision-centric tasks such as style transfer, a challenging area for current models. Using ISG-Bench, we demonstrate that recent unified vision-language models perform poorly on generating interleaved content. While compositional approaches that combine separate language and image models show a 111% improvement over unified models at the holistic level, their performance remains suboptimal at both block and image levels. To facilitate future work, we develop ISG-Agent, a baseline agent employing a *"plan-execute-refine"* pipeline to invoke tools, achieving a 122% performance improvement.

Recently, Multimodal Large Language Models (MLLMs) have been used as agents to control keyboard and mouse inputs by directly perceiving the Graphical User Interface (GUI) and generating corresponding commands. However, current agents primarily demonstrate strong understanding capabilities in static environments and are mainly applied to relatively simple domains, such as Web or mobile interfaces. We argue that a robust GUI agent should be capable of perceiving temporal information on the GUI, including dynamic Web content and multi-step tasks. Additionally, it should possess a comprehensive understanding of various GUI scenarios, including desktop software and multi-window interactions. To this end, this paper introduces a new dataset, termed GUI-World, which features meticulously crafted Human-MLLM annotations, extensively covering six GUI scenarios and eight types of GUI-oriented questions in three formats. We evaluate the capabilities of current state-of-the-art MLLMs, including Image LLMs and Video LLMs, in understanding various types of GUI content, especially dynamic and sequential content. Our findings reveal that current models struggle with dynamic GUI content without manually annotated keyframes or operation history. On the other hand, Video LLMs fall short in all GUI-oriented tasks given the sparse GUI video dataset. Therefore, we take the initial step of leveraging a fine-tuned Video LLM, GUI-Vid, as a GUI-oriented assistant, demonstrating an improved understanding of various GUI tasks. However, due to the limitations in the performance of base LLMs, we conclude that using video LLMs as GUI agents remains a significant challenge. We believe our work provides valuable insights for future research in dynamic GUI content understanding. All the dataset and code are publicly available at: https://gui-world.github.io.

LiDAR-based 3D perception plays a pivotal role in autonomous driving systems, which is a crucial component facilitating obstacle avoidance for vehicles, posing potential security risks in daily usage. Previous LiDAR-based adversarial attacks against autonomous driving models generally focus on the elimination or modification of specific objects in the point cloud. However, such attacks severely rely on the prior knowledge of the location of the target object and can only have limited adversarial impacts within certain local detection results. Instead, in this paper, we develop a novel LiDAR-based 3D adversarial attack from a more comprehensive perspective, which adaptively learns to impose global hazards without choosing particular local point sets of genuine obstacles, while enhancing the effectiveness, imperceptibility, and practicality of the perturbed LiDAR point clouds. We conduct extensive experiments on the large self-driving dataset Waymo to demonstrate the effectiveness of our attack against four object detection models (CenterPoint, PV-RCNN, DSVT, VoxelNext).

This paper proposes a novel method for multi-lane convoy formation control that uses large language models (LLMs) to tackle coordination challenges in dynamic highway environments. Each connected and autonomous vehicle in the convoy uses a knowledge-driven approach to make real-time adaptive decisions based on various scenarios. Our method enables vehicles to dynamically perform tasks, including obstacle avoidance, convoy joining/leaving, and escort formation switching, all while maintaining the overall convoy structure. We design a Interlaced formation control strategy based on locally dynamic distributed graphs, ensuring the convoy remains stable and flexible. We conduct extensive experiments in the SUMO simulation platform across multiple traffic scenarios, and the results demonstrate that the proposed method is effective, robust, and adaptable to dynamic environments. The code is available at: https://github.com/chuduanfeng/ConvoyLLM.

Large language models (LLMs) have demonstrated exceptional performance across a wide variety of domains. Nonetheless, generalist LLMs continue to fall short in reasoning tasks necessitating specialized knowledge, e.g., emotional sociology and medicine. Prior investigations into specialized LLMs focused on domain-specific training, which entails substantial efforts in domain data acquisition and model parameter fine-tuning. To address these challenges, this paper proposes the Way-to-Specialist (WTS) framework, which synergizes retrieval-augmented generation with knowledge graphs (KGs) to enhance the specialized capability of LLMs in the absence of specialized training. In distinction to existing paradigms that merely utilize external knowledge from general KGs or static domain KGs to prompt LLM for enhanced domain-specific reasoning, WTS proposes an innovative ''LLM↻KG'' paradigm, which achieves bidirectional enhancement between specialized LLM and domain knowledge graph (DKG). The proposed paradigm encompasses two closely coupled components: the DKG-Augmented LLM and the LLM-Assisted DKG Evolution. The former retrieves question-relevant domain knowledge from DKG and uses it to prompt LLM to enhance the reasoning capability for domain-specific tasks; the latter leverages LLM to generate new domain knowledge from processed tasks and use it to evolve DKG. WTS closes the loop between DKG-Augmented LLM and LLM-Assisted DKG Evolution, enabling continuous improvement in the domain specialization as it progressively answers and learns from domain-specific questions. We validate the performance of WTS on 7 datasets (e.g., TweetQA, ChatDoctor5k) spanning 6 domains, e.g., emotional sociology, medical, ect. The experimental results show that WTS surpasses the previous SOTA in 5 specialized domains, and achieves a maximum performance improvement of 11.3%.

Medical image segmentation is crucial for clinical decision-making, treatment planning, and disease tracking. Nonetheless, it confronts two significant challenges: the presence of ''soft boundaries'' between the foreground and background exacerbated by poor illumination and low contrast, and the misleading co-occurrence of salient and non-salient objects during the training phase, which complicates the model's accuracy in distinguishing relevant features. To overcome these challenges, we introduce RoDeCon-Net, a novel framework engineered to enhance medical image segmentation. RoDeCon-Net incorporates a Feature Decoupling Unit (FDU) that dynamically separates encoded features into foreground, background, and uncertain regions, using advanced attention mechanisms to refine feature distinction and reduce uncertainty. Additionally, our Contrast-driven Feature Alignment Unit (CFAU) and Cross-layer Feature Cascade Unit (CFCU) synergize to reinforce feature contrasts and promote effective multi-level feature fusion, thus improving the detection of salient objects amidst complex backgrounds and handling various object scales within images. Comprehensive evaluations of RoDeCon-Net on five diverse medical image datasets validate its superior performance and versatility, showcasing its potential to set new benchmarks in medical image segmentation. Our code is available on https://github.com/ILoveACM-MM/RoDeCon-Net.

Although Large Vision-Language Models (LVLMs) exhibit impressive multimodal capabilities, their vulnerability to adversarial examples has raised serious security concerns. Existing LVLM attackers simply optimize adversarial images that easily overfit a certain model/prompt, making them ineffective once they are transferred to attack a different model/prompt. Motivated by this research gap, this paper aims to develop a more powerful attack that is transferable to black-box LVLM models of different structures and task-aware prompts of different semantics. Specifically, we introduce a new perspective of information theory to investigate LVLMs' transferable characteristics by exploring the relative dependence between outputs of the LVLM model and input adversarial samples. Our empirical observations suggest that enlarging/decreasing the mutual information between outputs and the disentangled adversarial/benign patterns of input images helps to generate more agnostic perturbations for misleading LVLMs' perception with better transferability. In particular, we formulate the complicated calculation of information gain as an estimation problem and incorporate such informative constraints into the adversarial learning process. Extensive experiments on various LVLM models/prompts demonstrate our significant transfer-attack performance.

Self-correction of large language models (LLMs) emerges as a critical component for enhancing their reasoning performance. Although various self-correction methods have been proposed, a comprehensive evaluation of these methods remains largely unexplored, and the question of whether LLMs can truly correct themselves is a matter of significant interest and concern. In this study, we introduce CorrectBench, a benchmark developed to evaluate the effectiveness of self-correction strategies, including intrinsic, external, and fine-tuned approaches, across three tasks: commonsense reasoning, mathematical reasoning, and code generation. Our findings reveal that: 1) Self-correction methods can improve accuracy, especially for complex reasoning tasks; 2) Mixing different self-correction strategies yields further improvements, though it reduces efficiency; 3) Reasoning LLMs (e.g., DeepSeek-V3) have limited optimization under additional self-correction methods and have high time costs. Interestingly, a comparatively simple chain-of-thought (CoT) baseline demonstrates competitive accuracy and efficiency. These results underscore the potential of self-correction to enhance LLM's reasoning performance while highlighting the ongoing challenge of improving their efficiency. Consequently, we advocate for further research focused on optimizing the balance between reasoning capabilities and operational efficiency.

Although Multimodal Large Language Models (MLLMs) have demonstrated remarkable achievements in recent years, they remain vulnerable to adversarial examples that result in harmful responses. Existing attacks typically focus on optimizing adversarial perturbations for a certain multimodal image-prompt pair or fixed training dataset, which often leads to overfitting. Consequently, these perturbations fail to remain malicious once transferred to attack unseen image-prompt pairs, suffering from significant resource costs to cover the diverse multimodal inputs in complicated real-world scenarios. To alleviate this issue, this paper proposes a novel adversarial attack on MLLMs based on distribution approximation theory, which models the potential image-prompt input distribution and adds the same distribution-fitting adversarial perturbation on multimodal input pairs to achieve effective cross-image/prompt transfer attacks. Specifically, we exploit the Laplace approximation to model the Gaussian distribution of the image and prompt inputs for the MLLM, deriving an estimate of the mean and covariance parameters. By sampling from this approximated distribution with Monte Carlo mechanism, we efficiently optimize and fit a single input‑agnostic perturbation over diverse image‑prompt pairs, yielding strong universality and transferability. Extensive experiments are conducted to verify the strong adversarial capabilities of our proposed attack against prevalent MLLMs spanning a spectrum of images/prompts.

Understanding the internal mechanisms of transformer-based language models remains challenging. Mechanistic interpretability based on circuit discovery aims to reverse engineer neural networks by analyzing their internal processes at the level of computational subgraphs. In this paper, we revisit existing gradient-based circuit identification methods and find that their performance is either affected by the zero-gradient problem or saturation effects, where edge attribution scores become insensitive to input changes, resulting in noisy and unreliable attribution evaluations for circuit components. To address the saturation effect, we propose Edge Attribution Patching with GradPath (EAP-GP), EAP-GP introduces an integration path, starting from the input and adaptively following the direction of the difference between the gradients of corrupted and clean inputs to avoid the saturated region. This approach enhances attribution reliability and improves the faithfulness of circuit identification. We evaluate EAP-GP on 6 datasets using GPT-2 Small, GPT-2 Medium, and GPT-2 XL. Experimental results demonstrate that EAP-GP outperforms existing methods in circuit faithfulness, achieving improvements up to 17.7\%. Comparisons with manually annotated ground-truth circuits demonstrate that EAP-GP achieves precision and recall comparable to or better than previous approaches, highlighting its effectiveness in identifying accurate circuits.

Vision-Language-Action (VLA) models have advanced robotic control by enabling end-to-end decision-making directly from multimodal inputs. However, their tightly coupled architectures expose novel security vulnerabilities. Unlike traditional adversarial perturbations, backdoor attacks represent a stealthier, persistent, and practically significant threat—particularly under the emerging Training-as-a-Service paradigm—but remain largely unexplored in the context of VLA models. To address this gap, we propose **BadVLA**, a backdoor attack method based on Objective-Decoupled Optimization, which for the first time exposes the backdoor vulnerabilities of VLA models. Specifically, it consists of a two-stage process: (1) explicit feature-space separation to isolate trigger representations from benign inputs, and (2) conditional control deviations that activate only in the presence of the trigger, while preserving clean-task performance. Empirical results on multiple VLA benchmarks demonstrate that BadVLA consistently achieves near-100\% attack success rates with minimal impact on clean task accuracy. Further analyses confirm its robustness against common input perturbations, task transfers, and model fine-tuning, underscoring critical security vulnerabilities in current VLA deployments. Our work offers the first systematic investigation of backdoor vulnerabilities in VLA models, highlighting an urgent need for secure and trustworthy embodied model design practices.

Nowadays, as human activities are shifting to social media, fake news detection has been a crucial problem. Existing methods ignore the classification difference in online news and cannot take full advantage of multi-classification knowledges. For example, when coping with a post “A mouse is frightened by a cat,” a model that learns “computer” knowledge tends to misunderstand “mouse” and give a fake label, but a model that learns “animal” knowledge tends to give a true label. Therefore, this research proposes a multi-classification division-aggregation framework to detect fake news, namedCKA, which innovatively learns classification knowledges during training stages and aggregates them during prediction stages. It consists of three main components: a news characterizer, an ensemble coordinator, and a truth predictor. The news characterizer is responsible for extracting news features and obtaining news classifications. Cooperating with the news characterizer, the ensemble coordinator generates classification-specifical models for the maximum reservation of classification knowledges during the training stage, where each classification-specifical model maximizes the detection performance of fake news on corresponding news classifications. Further, to aggregate the classification knowledges during the prediction stage, the truth predictor uses the truth discovery technology to aggregate the predictions from different classification-specifical models based on reliability evaluation of classification-specifical models. Extensive experiments prove that our proposedCKAoutperforms state-of-the-art baselines in fake news detection.

With the significant development of large models in recent years, large vision-language models (LVLMs) have demonstrated remarkable capabilities across a wide range of multimodal understanding and reasoning tasks. Compared with traditional large language models (LLMs), LVLMs present great potential and challenges due to their closer proximity to the multiresource real-world applications and the complexity of multimodal processing. However, the vulnerability of LVLMs is relatively underexplored, posing potential security risks in the daily use of LVLM applications. In this article, we provide a comprehensive review of the various forms of existing LVLM attacks. Specifically, we first introduce the background of attacks targeting LVLMs, including the attack preliminary, attack challenges, and attack resources. Then, we systematically review the development of LVLM attack methods, such as adversarial attacks that manipulate model outputs, jailbreak attacks that exploit model vulnerabilities for unauthorized actions, prompt injection attacks that engineer the prompt type and pattern, and data poisoning that affects model training. Finally, we discuss promising future research directions in LVLM attacks. We believe that our survey provides insights into the current landscape of LVLM vulnerabilities, inspiring more researchers to explore and mitigate potential safety issues in LVLM developments.

In this paper, we study a connected submodular function maximization problem, which arises from many applications including deploying UAV networks to serve users and placing sensors to cover Points of Interest (PoIs). Specifically, given a budget K, the problem is to find a subset S with K nodes from a graph G, so that a given submodular function$f(S)$on S is maximized and the induced subgraph$G[S]$by the nodes in S is connected, where the submodular function f can be used to model many practical application problems, such as the number of users within different service areas of the deployed UAVs in S, the sum of data rates of users served by the UAVs, the number of covered PoIs by placed sensors, etc. We then propose a novel$\frac {1-1/e}{2h+2}$-approximation algorithm for the problem, improving the best approximation ratio$\frac {1-1/e}{2h+3}$for the problem so far, through estimating a novel upper bound on the problem and designing a smart graph decomposition technique, where e is the base of the natural logarithm, h is a parameter that depends on the problem and its typical value is 2. In addition, when$h=2$, the algorithm approximation ratio is at least$\frac {1-1/e}{5}$and may be as large as 1 in some special cases when$K\le 23$, and is no less than$\frac {1-1/e}{6}$when$K\ge 24$, compared with the current best approximation ratio$\frac {1-1/e}{7}\left ({{=\frac {1-1/e}{2h+3}}}\right)$for the problem. Finally, experimental results in the application of deploying a UAV network demonstrate that, the number of users within the service area of the deployed UAV network by the proposed algorithm is up to 7.5% larger than those by existing algorithms, and the throughput of the deployed UAV network by the proposed algorithm is up to 9.7% larger than those by the algorithms. Furthermore, the empirical approximation ratio of the proposed algorithm is between 0.7 and 0.99, which is close to the theoretical maximum value one.

Sequential Recommendation plays a significant role in daily recommendation systems, such as e-commerce platforms like Amazon and Taobao. However, even with the advent of large models, these platforms often face sparse issues in the historical browsing records of individual users due to new users joining or the introduction of new products. As a result, existing sequence recommendation algorithms may not perform well. To address this, sequence-based data augmentation methods have garnered attention. Existing sequence enhancement methods typically rely on augmenting existing data, employing techniques like cropping, masking prediction, random reordering, and random replacement of the original sequence. While these methods have shown improvements, they often overlook the exploration of the deep embedding space of the sequence. To tackle these challenges, we propose a Sparse Enhanced Network (SparseEnNet), which is a robust adversarial generation method. SparseEnNet aims to fully explore the hidden space in sequence recommendation, generating more robust enhanced items. Additionally, we adopt an adversarial generation method, allowing the model to differentiate between data augmentation categories and achieve better prediction performance for the next item in the sequence. Experiments have demonstrated that our method achieves a remarkable 4-14% improvement over existing methods when evaluated on the real-world datasets. (https://github.com/junyachen/SparseEnNet)

Given an untrimmed video and a sentence query, video moment retrieval using language (VMR) aims to locate a target query-relevant moment. Since the untrimmed video is overlong, almost all existing VMR methods first sparsely down-sample each untrimmed video into multiple fixed-length video clips and then conduct multi-modal interactions with the query feature and expensive clip features for reasoning, which is infeasible for long real-world videos that span hours. Since the video is downsampled into fixed-length clips, some query-related frames may be filtered out, which will blur the specific boundary of the target moment, take the adjacent irrelevant frames as new boundaries, easily leading to cross-modal misalignment and introducing both boundary-bias and reasoning-bias. To this end, in this paper, we propose an efficient approach, SpotVMR, to trim the query-relevant clip. Besides, our proposed SpotVMR can serve as plug-and-play module, which achieves efficiency for state-of-the-art VMR methods while maintaining good retrieval performance. Especially, we first design a novel clip search model that learns to identify promising video regions to search conditioned on the language query. Then, we introduce a set of low-cost semantic indexing features to capture the context of objects and interactions that suggest where to search the query-relevant moment. Also, the distillation loss is utilized to address the optimization issues arising from end-to-end joint training of the clip selector and VMR model. Extensive experiments on three challenging datasets demonstrate its effectiveness.

Temporal sentence localization (TSL) aims to localize a target segment in a video according to a given sentence query. Though respectable works have made decent achievements in this task, they severely rely on abundant yet expensive manual annotations for training. Moreover, these trained data-dependent models usually can not generalize well to unseen scenarios because of the inherent domain shift. To facilitate this issue, in this paper, we target another more practical but challenging setting: unsupervised domain adaptative temporal sentence localization (UDA-TSL), which explores whether the localization knowledge can be transferred from a fully-annotated data domain (source domain) to a new unannotated data domain (target domain). Particularly, we propose an effective and novel baseline for UDA-TSL to bridge the multi-modal gap across different domains and learn the potential correspondence between the video-query pairs in target domain. We first develop separate modality-specific domain adaptation modules to smoothly balance the minimization of the domain shifts in cross-dataset video and query domains. Then, to fully exploit the semantic correspondence of both modalities in target domain for unsupervised localization, we devise a mutual information learning module to adaptively align the video-query pairs which are more likely to be relevant in target domain, leading to more truly aligned target pairs and ensuring the discriminability of target features. In this way, our model can learn domain-invariant and semantic-aligned cross-modal representations. Three sets of migration experiments show that our model achieves competitive performance compared to existing methods.

Graph neural networks (GNNs) have recently been shown to be vulnerable to adversarial attacks, where slight perturbations in the graph structure can lead to erroneous predictions. However, current robust models for defending against such attacks inherit the transductive limitations of graph convolutional networks (GCNs). As a result, they are constrained by fixed structures and do not naturally generalize to unseen nodes. Here, we discover that transductive GCNs inherently possess a distillable robustness, achieved through a wave-induced resonance process. Based on this, we foster this resonance to facilitate inductive and robust learning. Specifically, we first prove that the signal formed by GCN-driven message passing (MP) is equivalent to the edge-based Laplacian wave, where, within a wave system, resonance can naturally emerge between the signal and its transmitting medium. This resonance provides inherent resistance to malicious perturbations inflicted on the signal system. We then prove that merely three MP iterations within GCNs can induce signal resonance between nodes and edges, manifesting as a coupling between nodes and their distillable surrounding local subgraph. Consequently, we present Graph Resonance-fostering Network (GRN) to foster this resonance via learning node representations from their distilled resonating subgraphs. By capturing the edge-transmitted signals within this subgraph and integrating them with the node signal, GRN embeds these combined signals into the central node's representation. This node-wise embedding approach allows for generalization to unseen nodes. We validate our theoretical findings with experiments, and demonstrate that GRN generalizes robustness to unseen nodes, whilst maintaining state-of-the-art classification accuracy on perturbed graphs. Appendices can be found on arXiv version: https://arxiv.org/abs/2312.08651

Multi-agent perception (MAP) allows autonomous systems to understand complex environments by interpreting data from multiple sources. This paper investigates intermediate collaboration for MAP with a specific focus on exploring "good" properties of collaborative view (i.e., post-collaboration feature) and its underlying relationship to individual views (i.e., pre-collaboration features), which were treated as an opaque procedure by most existing works. We propose a novel framework named CMiMC (Contrastive Mutual Information Maximization for Collaborative Perception) for intermediate collaboration. The core philosophy of CMiMC is to preserve discriminative information of individual views in the collaborative view by maximizing mutual information between pre- and post-collaboration features while enhancing the efficacy of collaborative views by minimizing the loss function of downstream tasks. In particular, we define multi-view mutual information (MVMI) for intermediate collaboration that evaluates correlations between collaborative views and individual views on both global and local scales. We establish CMiMNet based on multi-view contrastive learning to realize estimation and maximization of MVMI, which assists the training of a collaborative encoder for voxel-level feature fusion. We evaluate CMiMC on V2X-Sim 1.0, and it improves the SOTA average precision by 3.08% and 4.44% at 0.5 and 0.7 IoU (Intersection-over-Union) thresholds, respectively. In addition, CMiMC can reduce communication volume to 1/32 while achieving performance comparable to SOTA. Code and Appendix are released at https://github.com/77SWF/CMiMC.

Adversarial attacks on 3D point clouds often exhibit unsatisfactory imperceptibility, which primarily stems from the disregard for manifold-aware distortion, i.e., distortion of the underlying 2-manifold surfaces. In this paper, we develop novel manifold constraints to reduce such distortion, aiming to enhance the imperceptibility of adversarial attacks on 3D point clouds. Specifically, we construct a bijective manifold mapping between point clouds and a simple parameter shape using an invertible auto-encoder. Consequently, manifold-aware distortion during attacks can be captured within the parameter space. By enforcing manifold constraints that preserve local properties of the parameter shape, manifold-aware distortion is effectively mitigated, ultimately leading to enhanced imperceptibility. Extensive experiments demonstrate that integrating manifold constraints into conventional adversarial attack solutions yields superior imperceptibility, outperforming the state-of-the-art methods.

Zhaochen Su, Juntao Li, Jun Zhang, Tong Zhu, Xiaoye Qu, Pan Zhou, Yan Bowen, Yu Cheng, Min Zhang. Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). 2024.

With the Mamba framework widely applied, the state-space model has yielded outstanding results in the field of computer vision. Nevertheless, the superiority of the model to its counterparts, namely, CNN-based or Transformer-based models, is limited, because it faces large challenges in local region feature extraction resulting from the deficient positional awareness and the disproportional emphasis on posterior tokens in its pre-defined scanning schedules. Besides, a majority of the current Mamba-based models usually fail to take into account the advantages of the integration of multi-visual encoding strategies. Against this background, we propose a novel DoubleU-Net framework with multiple visual encoding strategies and a local-based scanning mechanism. The comparative & ablation experiments with the current SOTA methods verify the superiority or competitiveness of the DUMFNet. For reproduction, the implementation codes can be checked out at https://github.com/Panpz202006/DUMFNet.

LLM-as-a-Judge uses a large language model (LLM) to select the best response from a set of candidates for a given question. LLM-as-a-Judge has many applications such as LLM-powered search, reinforcement learning with AI feedback (RLAIF), and tool selection. In this work, we propose JudgeDeceiver, an optimization-based prompt injection attack to LLM-as-a-Judge. JudgeDeceiver injects a carefully crafted sequence into an attacker-controlled candidate response such that LLM-as-a-Judge selects the candidate response for an attacker-chosen question no matter what other candidate responses are. Specifically, we formulate finding such sequence as an optimization problem and propose a gradient based method to approximately solve it. Our extensive evaluation shows that JudgeDeceive is highly effective, and is much more effective than existing prompt injection attacks that manually craft the injected sequences and jailbreak attacks when extended to our problem. We also show the effectiveness of JudgeDeceiver in three case studies, i.e., LLM-powered search, RLAIF, and tool selection. Moreover, we consider defenses including known-answer detection, perplexity detection, and perplexity windowed detection. Our results show these defenses are insufficient, highlighting the urgent need for developing new defense strategies.

This paper addresses the task of temporal activity localization (TAL). Although recent works have made significant progress in TAL research, almost all of them implicitly assume that the dense frame-level correspondences in each video-query pair are correctly annotated. However, in reality, such an assumption is extremely expensive and even impossible to satisfy due to subjective labeling. To alleviate this issue, in this paper, we explore a new TAL setting termed Noisy Temporal activity localization (NTAL), where a TAL model should be robust to the mixed training data with noisy moment boundaries. Inspired by the memorization effect of neural networks, we propose a novel method called Co-Teaching Regularizer (CTR) for NTAL. Specifically, we first learn a Gaussian Mixture Model to divide the mixed training data into preliminary clean and noisy subsets. Subsequently, we refine the labels of the two subsets by an adaptive prediction function so that their true positive and false positive samples could be identified. To avoid single model being prone to its mistakes learned by the mixed data, we adopt a co-teaching paradigm, which utilizes two models sharing the same framework to teach each other for robust learning. A curriculum strategy is further introduced to gradually learn the moment confidence from easy to hard. Experiments on three datasets demonstrate that our CTR is significantly more robust to the noisy training data compared to the existing methods.

Text-to-image diffusion models allow seamless generation of personalized images from scant reference photos. Yet, these tools, in the wrong hands, can fabricate misleading or harmful content, endangering individuals. To address this problem, existing poisoning-based approaches perturb user images in an imperceptible way to render them “unlearnable” from malicious uses. We identify two limitations of these defending approaches: i) sub-optimal due to the hand-crafted heuristics for solving the intractable bilevel optimization and ii) lack of robustness against simple data transformations like Gaussian filtering. To solve these challenges, we propose MetaCloak, which solves the bi-level poisoning problem with a meta-learning framework with an additional transformation sampling process to craft transferable and robust perturbation. Specifically, we employ a pool of surrogate diffusion models to craft transferable and model-agnostic perturbation. Furthermore, by incorporating an additional transformation process, we design a simple denoising-error maximization loss that is sufficient for causing transformation-robust semantic distortion and degradation in a personalized generation. Extensive experiments on the VGGFace2 and CelebA-HQ datasets show that MetaCloak outperforms existing approaches. Notably, MetaCloak can successfully fool online training services like Replicate, in a black-box manner, demonstrating the effectiveness of Meta Cloak in real-world scenarios. Our code is available at https://github.com/liuyixin-louis/MetaCloak.

Backdoor attacks have been well-studied in visible light object detection (VLOD) in recent years. However, VLOD can not effectively work in dark and temperature-sensitive scenarios. Instead, thermal infrared object detection (TIOD) is the most accessible and practical in such environments. In this paper, our team is the first to investigate the security vulnerabilities associated with TIOD in the context of backdoor attacks, spanning both the digital and physical realms. We introduce two novel types of backdoor attacks on TIOD, each offering unique capabilities: Object-affecting Attack and Range-affecting Attack. We conduct a comprehensive analysis of key factors influencing trigger design, which include temperature, size, material, and concealment. These factors, especially temperature, significantly impact the efficacy of backdoor attacks on TIOD. A thorough understanding of these factors will serve as a foundation for designing physical triggers and temperature controlling experiments. Our study includes extensive experiments conducted in both digital and physical environments. In the digital realm, we evaluate our approach using benchmark datasets for TIOD, achieving an Attack Success Rate (ASR) of up to 98.21%. In the physical realm, we test our approach in two real-world settings: a traffic intersection and a parking lot, using a thermal infrared camera. Here, we attain an ASR of up to 98.38%.

In this paper, we study a connected submodular function maximization problem, which arises from many applications including deploying UAV networks to serve users and placing sensors to cover Points of Interest (PoIs). Specifically, given a budget$K$, the problem is to find a subset$S$with$K$nodes from a graph$G$so that a given submodular function$f (S)$on$S$is maximized while the induced subgraph$G[S]$by the nodes in$S$is connected, where the submodular function$f$can be used to model many practical application problems, such as the number of users within different service areas of the deployed UAVs in$S$, the sum of data rates of users served by the UAVs, the number of covered PoIs by placed sensors, etc. We then propose a novel$\frac{1-1/e}{2h+2}$-approximation algorithm for the problem, improving the best approximation ratio$\frac{1-1/e}{2h+3}$for the problem so far, through estimating a novel upper bound on the problem and designing a smart graph decomposition technique, where$e$is the base of the natural logarithm,$h$is a parameter depends on the problem and its typical value is 2. In addition. when$h= 2$, the algorithm approximation ratio is at least$\frac{1-1/e}{5}$and may be as large as 1 in some special cases when$K$≤21, and is no less than$\frac{1-1/e}{6}$when$K$≥ 22, compared with the current best approximation ratio$\frac{1-1/e}{7}(= \frac{1-1/e}{2h+3})$for the problem. We finally evaluate the algorithm performance in the application of deploying a UAV network. Experimental results demonstrate the number of users within the service area of the deployed UAV network by the proposed algorithm is up to 7.5% larger than those by existing algorithms, and its empirical approximation ratio is between 0.7 and 0.99, which is close to the theoretical maximum value one.

Large language models (LLMs) have garnered significant attention due to their impressive natural language processing (NLP) capabilities. Recently, many studies have focused on the tool utilization ability of LLMs. They primarily investigated how LLMs effectively collaborate with given specific tools. However, in scenarios where LLMs serve as intelligent agents, as seen in applications like AutoGPT and MetaGPT, LLMs are expected to engage in intricate decision-making processes that involve deciding whether to employ a tool and selecting the most suitable tool(s) from a collection of available tools to fulfill user requests. Therefore, in this paper, we introduce MetaTool, a benchmark designed to evaluate whether LLMs have tool usage awareness and can correctly choose tools. Specifically, we create a dataset called ToolE within the benchmark. This dataset contains various types of user queries in the form of prompts that trigger LLMs to use tools, including both single-tool and multi-tool scenarios. Subsequently, we set the tasks for both tool usage awareness and tool selection. We define four subtasks from different perspectives in tool selection, including tool selection with similar choices, tool selection in specific scenarios, tool selection with possible reliability issues, and multi-tool selection. We conduct experiments involving eight popular LLMs and find that the majority of them still struggle to effectively select tools, highlighting the existing gaps between LLMs and genuine intelligent agents. However, through the error analysis, we found there is still significant room for improvement. Finally, we conclude with insights for tool developers -- we strongly recommend that tool developers choose an appropriate rewrite model for generating new descriptions based on the downstream LLM the tool will apply to.

Multimodal Large Language Models (MLLMs) have gained significant attention recently, showing remarkable potential in artificial general intelligence. However, assessing the utility of MLLMs presents considerable challenges, primarily due to the absence multimodal benchmarks that align with human preferences. Drawing inspiration from the concept of LLM-as-a-Judge within LLMs, this paper introduces a novel benchmark, termed MLLM-as-a-Judge, to assess the ability of MLLMs in assisting judges across diverse modalities, encompassing three distinct tasks: Scoring Evaluation, Pair Comparison, and Batch Ranking. Our study reveals that, while MLLMs demonstrate remarkable human-like discernment in Pair Comparisons, there is a significant divergence from human preferences in Scoring Evaluation and Batch Ranking tasks. Furthermore, a closer examination reveals persistent challenges in the evaluative capacities of LLMs, including diverse biases, hallucinatory responses, and inconsistencies in judgment, even in advanced models such as GPT-4V. These findings emphasize the pressing need for enhancements and further research efforts to be undertaken before regarding MLLMs as fully reliable evaluators. In light of this, we advocate for additional efforts dedicated to supporting the continuous development within the domain of MLLM functioning as judges. The code and dataset are publicly available at our project homepage: https://mllm-judge.github.io/.

Modern machine learning applications increasingly demand greater computational resources for training large models. Decentralized training has emerged as an effective means to democratize this technology. However, the potential threats associated with this approach remain inadequately discussed, posing a hurdle to the development of decentralized training infrastructures. This paper aims to initiate discussion towards this end by exploring the robustness of decentralized training from three primary perspectives. Firstly, we articulate our position on establishing robust decentralized training by outlining potential threats and the corresponding countermeasures. Secondly, we illustrate a nascent poisoning attack targeting decentralized training frameworks, easily executable by malicious stages. To mitigate this security threat and ensure efficient training, we propose a robust training framework, integrating a 100% detection strategy and efficient training mechanisms. Finally, we demonstrate the severity of the proposed attack and the effectiveness of our robust training framework. This position paper emphasizes the urgency of exploring the robustness of decentralized training and proposes a feasible solution. The code is available at https://github.com/dcx001016/pipeline_attack.

With the development of depth sensors and 3D vision, the vulnerability of 3D point cloud models has garnered heightened concern. Almost all existing 3D attackers are deployed in the white-box setting, where they access the model details and directly optimize coordinate-wise noises to perturb 3D objects. However, realistic 3D applications would not share any model information (model parameters, gradients, etc.) with users. Although a few recent works try to explore the black-box attack, they still achieve limited attack success rates (ASR) and fail to generate high-quality adversarial samples. In this paper, we focus on designing a transfer-based black-box attack method, called Transferable Frequency-aware 3D GAN, to delve into achieving a high black-box ASR by improving the adversarial transferability while making the adversarial samples more imperceptible. Considering that the 3D imperceptibility depends on whether the shape of the object is distorted, we utilize the spectral tool with the GAN design to explicitly perceive and preserve the 3D geometric structures. Specifically, we design the Graph Fourier Transform (GFT) encoding layer in the GAN generator to extract the geometries as guidance, and develop a corresponding Inverse-GFT decoding layer to decode latent features with this guidance to reconstruct high-quality adversarial samples. To further improve the transferability, we develop a dual learning scheme of discriminator from both frequency and feature perspectives to constrain the generator via adversarial learning. Finally, imperceptible and transferable perturbations are rapidly generated by our proposed attack. Experimental results demonstrate that our attack method achieves the highest transfer ASR while exhibiting stronger imperceptibility.

Video Moment Retrieval (VMR) targets to retrieve the specific moment corresponding to a sentence query from an untrimmed video. Although recent respectable works have made remarkable progress in this task, they implicitly are rooted in the closed-set assumption that all the given queries as video-relevant. Given an OOD query in open-set scenarios, they still utilize it for wrong retrieval, which might lead to irrecoverable losses in high-risk scenarios, e.g., criminal activity detection. To this end, we creatively explore a brand-new VMR setting termed Open-Set Video Moment Retrieval (OS-VMR), where we should not only retrieve the precise moments based on ID query, but also reject OOD queries. In this paper, we make the first attempt to step toward OS-VMR and propose a novel model OpenVMR, which first distinguishes ID and OOD queries based on the normalizing flow technology, and then conducts moment retrieval based on ID queries. Specifically, we first learn the ID distribution by constructing a normalizing flow, and assume the ID query distribution obeys the multi-variate Gaussian distribution. Then, we introduce an uncertainty score to search the ID-OOD separating boundary. After that, we refine the ID-OOD boundary by pulling together ID query features. Besides, video-query matching and frame-query matching are designed for coarse-grained and fine-grained cross-modal interaction, respectively. Finally, a positive-unlabeled learning module is introduced for moment retrieval. Experimental results on three VMR datasets show the effectiveness of our OpenVMR.

RGB-Thermal Salient Object Detection (RGBT-SOD) plays a critical role in complex scene recognition applications, such as autonomous driving. However, security research in this domain is still in its infancy. This paper presents the first backdoor attack on RGBT-SOD systems, generating saliency maps on triggered inputs that depict non-existent salient objects chosen by the attacker or falsely mark an entire image as fully salient or entirely non-salient. We uncover that triggers have an influence range for generating non-existent salient objects, supported by a theoretical analysis. Extensive experiments show the effectiveness of our attack in both digital and physical-world scenarios. Notably, our dual-modality backdoor attack achieves an Attack Success Rate (ASR) of 86.72% with only five pairs of poisoned images in model training. After investigating potential countermeasures, we find them inadequate in mitigating our attacks, highlighting the urgent need for robust defenses against sophisticated backdoor attacks in RGBT-SOD systems.

Large Vision-Language Models (LVLMs) have demonstrated remarkable capabilities across a wide range of multimodal understanding tasks. Nevertheless, these models are susceptible to adversarial examples. In real-world applications, existing LVLM attackers generally rely on the detailed prior knowledge of the model to generate effective perturbations. Moreover, these attacks are task-specific, leading to significant costs for designing perturbation. Motivated by the research gap and practical demands, in this paper, we make the first attempt to build a universal attacker against real-world LVLMs, focusing on two critical aspects: (i) restricting access to only the LVLM inputs and outputs. (ii) devising a universal adversarial patch, which is task-agnostic and can deceive any LVLM-driven task when applied to various inputs. Specifically, we start by initializing the location and the pattern of the adversarial patch through random sampling, guided by the semantic distance between their output and the target label. Subsequently, we maintain a consistent patch location while refining the pattern to enhance semantic resemblance to the target. In particular, our approach incorporates a diverse set of LVLM task inputs as query samples to approximate the patch gradient, capitalizing on the importance of distinct inputs. In this way, the optimized patch is universally adversarial against different tasks and prompts, leveraging solely gradient estimates queried from the model. Extensive experiments are conducted to verify the strong universal adversarial capabilities of our proposed attack with prevalent LVLMs including LLaVA, MiniGPT-4, Flamingo, and BLIP-2, spanning a spectrum of tasks, all achieved without delving into the details of the model structures.

Graph neural network (GNN), the mainstream method to learn on graph data, is vulnerable to graph evasion attacks, where an attacker slightly perturbing the graph structure can fool trained GNN models. Existing work has at least one of the following drawbacks: 1) limited to directly attack two-layer GNNs; 2) inefficient; and 3) impractical, as they need to know full or part of GNN model parameters.

This paper proposes Meta Parallel Graph Neural Network (MPGNN) to establish a scalable Network Intrusion Detection System (NIDS) for large-scale Internet of Things (IoT) networks. MPGNN leverages a meta-learning framework to optimize the parallelism of GNN-based NIDS. The core of MPGNN is a coalition formation policy that generates meta-knowledge for partitioning a massive graph into multiple coalitions/subgraphs in a way that maximizes the performance and efficiency of parallel coalitional NIDSs. We propose an offline reinforcement learning algorithm, called Graph-Embedded Adversarially Trained Actor-Critic (G-ATAC), to learn a coalition formation policy that jointly optimizes intrusion detection accuracy, communication overheads, and computational complexities of coalitional NIDSs. In particular, G-ATAC learns to capture the temporal dependencies of network states and coalition formation decisions over offline data, eliminating the need for expensive online interactions with large IoT networks. Given generated coalitions, MPGNN employs E-GraphSAGE to establish coalitional NIDSs which then collaborate via ensemble prediction to accomplish intrusion detection for the entire network. We evaluate MPGNN on two real-world datasets. The experimental results demonstrate the superiority of our method with substantial improvements in F1 score, surpassing the state-of-the-art methods by 0.38 and 0.29 for the respective datasets. Compared to the centralized NIDS, MPGNN reduces the training time of NIDS by 41.63% and 22.11%, while maintaining an intrusion detection performance comparable to centralized NIDS.

Low Earth Orbit (LEO) satellite networks provide global connectivity but are vulnerable to security threats such as link flooding attacks. To defend against such attacks, stateof-the-art approaches employ SDN to acquire a global view of the network, enabling the detection and mitigation of malicious traffic. However, in LEO constellation networks, the distributed nature of satellites across a large spatial scale introduces significant latency in both satellite-to-ground and inter-satellite links, with latency reaching up to tens of milliseconds, while attack traffic dynamically adapts within sub-milliseconds. As a result, existing defense systems face challenges in countering these attacks effectively due to the increased reaction time caused by link latency. In this paper, we leverage programmable switches to build a real-time defense system against link flooding attacks (LFA) in LEO constellation networks. To achieve this, we analyze the practical constraints encountered in the deployment of LFA attacks against state-of-the-art LEO satellite systems. We observe that despite the ability of bots to initiate attack traffic from any location worldwide, an anomalous distribution of flow rate on the affected links can still be detected. We propose SatShield, an in-network defense system that filters out suspicious traffic (heavy flows) in the network and mitigates these threats by leveraging programmable packet scheduling. By using SatShield, we are able to achieve real-time identification and rate-limiting of attacks at line rate on a per-packet basis. We implement SatShield with P4 in a commercial programmable switch and evaluate it with real-world traffic traces. Our evaluation shows that SatShield autonomously identifies LFA attack flows and rapidly mitigates LFA attacks.

Federated continual learning (FCL) is emerging as a key technology for time-sensitive applications in highly adaptive environments including autonomous driving and industrial digital twin. Each FCL trains machine learning models using newly-generated datasets as soon as possible, to obtain a highly accurate machine learning model for new event predictions. Theage of data, defined as the time difference between the generation time of a dataset and the current time, is widely adopted as a key criterion to evaluate both timeline and quality of training. In this paper, we study the problem of age-aware FCL in a mobile edge computing (MEC) network. We not only investigate optimization techniques that optimize the data selection and aggregator placement for FCL but also implement a real system as a prototype for age-aware FCL. Specifically, we first propose an approximation algorithm with a provable approximation ratio for the age-aware data selection and aggregator placement problem for FCL with a single request. In real application scenarios, there are usually multiple FCL requests that require to train models, and delays in the MEC network are usually uncertain. We then study the problem of age-aware data selection and aggregator placement problem for FCL with uncertain delays and multiple requests, by devising an online learning algorithm with a bounded regret based on contextual bandits. We finally implement a prototype for FCL in an MEC network, with various heterogeneous user equipments (UEs) and cloudlets with different computing capabilities in the network. Experiment results show that the performance of the proposed algorithms outperform existing studies, by achieving 47% lower age of data and 12% higher model accuracy.

Federated learning (FL) has attracted increasing attention in recent years due to its data privacy preservation and great applicability to large-scale user scenarios. However, when FL faces numerous clients, it is inevitable to emerge the non-independent and identically distributed (non-iid) data between clients, which brings an enormous challenge for model training and performance analysis like convergence. Besides, due to the non-iid data, the participating clients of FL tend to be extremely heterogeneous so the number of samplings among clients causes a sampling variance problem, which induces a huge variation in convergence. More importantly, although FL can foster privacy security via locally retaining the training data, if local data is secret and sensitive, FL should have more powerful privacy protection to resist the cloud server or third party to infer private information from shared models or intermediate gradients. Facing the non-iid and privacy challenges, we propose a differential privacy (DP) based non-iid FL algorithm called DPNFL to jointly tackle these two issues. Specifically, motivated by the DP and its variants, we are the first to adopt the truncated concentrated differential privacy technique under the FL scenario to more tightly track end-to-end privacy loss, while requiring less noise injection for the same level of DP. To avoid the sampling variance problem, we enable the server to sample the partial clients uniformly without replacement, which also guarantees unbiased sampling. To further improve the algorithm performance, we also propose an adaptive version of DPNFL named AdDPNFL, which adopts the adaptive optimization on the server-side to simultaneously alleviate the impact of non-iid data and DP noise on model utility. Finally, we perform extensive experiments to validate the effectiveness and superiority of our algorithms.

Microservice (MS) structures a service application as a collection of independently deployable service modules, making it particularly suitable for delivering complex applications in distributed computing systems. This paper investigates MS architecture over Mobile Edge Computing (MEC) networks (hereafter referred to as EdgeMS) and studies an EdgeMS placement problem that aims to deploy MS modules over the MEC network in a manner that maximizes the reward of MS application providers. A novel algorithm called Dual-GNN Deep Deterministic Policy Gradient (DG-DDPG) is proposed to establish an intelligent EdgeMS placement policy for optimizing the location of MS modules and performing fractional computing resource allocation. DG-DDPG leverages the graph neural network (GNN) to comprehend the graph-structured information encapsulated in the MS application structure and MEC network. A dual-GNN core is constructed in DG-DDPG, one GNN for MS applications to distill knowledge from intricate connections between MS modules, and the other GNN for MEC networks to capture complicated interactions between edge sites when providing EdgeMS. DG-DDPG embeds the dual-GNN core in a DDPG-based reinforcement learning framework, which not only handles temporal dependencies between EdgeMS placement decisions for maximizing long-term reward but also supports continuous action space for enabling fractional resource allocation. In particular, the learning process of DG-DDPG is tailored to address hard constraints (i.e., computing capacity and MS application completeness) in the EdgeMS placement problem. We design constraint-based regularization terms and add them to the objective of DG-DDPG, which facilitates the identification of feasible placement decisions during learning. We carry out systematic experiments to evaluate the performance of DG-DDPG, and the results show that DG-DDPG outperforms state-of-the-art benchmarks in terms of reward, service delay and deployment cost.

The escalating proliferation of user generated contents such as videos and images are dominating the network traffic. The optimal strategy for mitigating backbone congestion and minimizing user request latency lies in prudent caching at edge stations within distributed autonomous networks, obviating the necessity to transmit data to the cloud. However, accurately caching content based on distributed autonomous networks requires elaborative collaboration between edge servers, which remains a great challenge, especially when content is highly dynamic and the storage resources of edge stations are limited. To tackle this challenge, this paper proposes a contextual bandit-based online caching algorithm for evaluating the optimal content hit rate reward, which can adapt to the constantly changing stream of emerging content. We build the content space, BS space, and a fine-grained space searching method to cache contents and corresponding edge stations. Furthermore, to perform collaborative caching and sharing between edges, we propose a federated autonomous multi-layer caching framework, whereby each server can locally learn the model for accurate caching and a synchronous mechanism is set up for global updating, further improving the hit rates. Finally, we perform theoretical proofs and simulations, demonstrating that our regret is sublinear and our caching algorithm outperforms several state-of-the-art algorithms.

To unveil the hidden value in the datasets of user equipments (UEs) while preserving user privacy, federated learning (FL) is emerging as a promising technique to train a machine learning model using the datasets of UEs locally without uploading the datasets to a central location. Customers require to train machine learning models based on different datasets of UEs, through issuing FL requests that are implemented by FL services in a mobile edge computing (MEC) network. A key challenge of enabling FL in MEC networks is how to minimize the energy consumption of implementing FL requests while guaranteeing the accuracy of machine learning models, given that the availabilities of UEs usually are uncertain. In this paper, we investigate the problem of energy minimization for FL in an MEC network with uncertain availabilities of UEs. We first consider the energy minimization problem for a single FL request in an MEC network. We then propose a novel optimization framework for the problem with a single FL request, which consists of (1) an online learning algorithm with a bounded regret for the UE selection, by considering various contexts (side information) that influence energy consumption; and (2) an approximation algorithm with an approximation ratio for the aggregator placement for a single FL request. We third deal with the problem with multiple FL requests, for which we devise an online learning algorithm with a bounded regret. We finally evaluate the performance of the proposed algorithms by extensive experiments. Experimental results show that the proposed algorithms outperform their counterparts by reducing at least 13% of the total energy consumption while achieving the same accuracy.

This article studies the multimedia problem of temporal sentence grounding (TSG), which aims to accurately determine the specific video segment in an untrimmed video according to a given sentence query. Traditional TSG methods mainly follow the top-down or bottom-up framework and are not end-to-end. They severely rely on time-consuming post-processing to refine the grounding results. Recently, some transformer-based approaches are proposed to efficiently and effectively model the fine-grained semantic alignment between video and query. Although these methods achieve significant performance to some extent, they equally take frames of the video and words of the query as transformer input for correlating, failing to capture their different levels of granularity with distinct semantics. To address this issue, in this article, we propose a novelHierarchicalLocal-GlobalTransformer (HLGT) to leverage this hierarchy information and model the interactions between different levels of granularity and different modalities for learning more fine-grained multi-modal representations. Specifically, we first split the video and query into individual clips and phrases to learn their local context (adjacent dependency) and global correlation (long-range dependency) via a temporal transformer. Then, a global-local transformer is introduced to learn the interactions between the local-level and global-level semantics for better multi-modal reasoning. Besides, we develop a new cross-modal cycle-consistency loss to enforce interaction between two modalities and encourage the semantic alignment between them. Finally, we design a brand-new cross-modal parallel transformer decoder to integrate the encoded visual and textual features for final grounding. Extensive experiments on three challenging datasets (ActivityNet Captions, Charades-STA and TACoS) show that our proposed HLGT achieves a new state-of-the-art performance, demonstrating its effectiveness and computational efficiency.

Temporal sentence grounding (TSG) aims to locate a semantically related segment of an untrimmed video guided by a sentence query. Since the untrimmed videos are too long, almost all existing TSG works first sparsely down-sample each video into a shorter video of a fixed length and then conduct multimodal interactions with the query sentence for reasoning. However, this video down-sampling process may introduce a challenging issue that confuses the latter grounding process: Due to the video down-sampling, some query-related frames may be filtered out; this process may remove the specific boundary frames of the target segment and take the adjacent irrelevant frames as new boundaries, easily leading to cross-modal misalignment and introducing both boundary-bias and reasoning-bias. Therefore, it is important to keep the grounding consistency (both temporal annotations and boundary predictions) between the original and the sampled videos. To this end, in this paper, we propose a novel Conditional Video Diffusion Network (CVDN) for TSG to learn extra visual semantics to enrich and refine the biased new boundaries, which enables soft-label boundary prediction for fine-grained frame-query reasoning. Specifically, we first construct a conditional video diffusion model which is separately trained to recover the consecutive semantics of the filtered frames between the adjacent sampled frames. Through the designed stochastic interval sampling strategies in the training process, this diffusion model is able to generate absent coherent semantics between the sparsely sampled frames and in turn enrich and refine them, benefiting the integral activity understanding for TSG. In this manner, the incorrect new boundaries will be refined to be closely correlated to the original boundary frames and contain sufficient query-related information, which is crucial for accurate segment prediction. Extensive experiments on three challenging datasets demonstrate the effectiveness of CVDN.

Video sentence grounding (VSG) is the task of identifying the segment of an untrimmed video that semantically corresponds to a given natural language query. While many existing methods extract frame-grained features using pre-trained 2D or 3D convolution networks, often fail to capture subtle differences between ambiguous adjacent frames. Although some recent approaches incorporate object-grained features using Faster R-CNN to capture more fine-grained details, they are still primarily based on feature enhancement and lack spatio-temporal modeling to explore the semantics of the core persons/objects. To solve the problem of modeling the core target's behavior, in this paper, we propose a new perspective for addressing the VSG task by tracking pivotal objects and activities to learn more fine-grained spatio-temporal features. Specifically, we introduce the Video Sentence Tracker with Memory Network and Masked Attention (VSTMM), which comprises a cross-modal targets generator for producing multi-modal templates and search space, a memory-based tracker for dynamically tracking multi-modal targets using a memory network to record targets' behaviors, a masked attention localizer which learns local shared features between frames and eliminates interference from long-term dependencies, resulting in improved accuracy when localizing the moment. To evaluate the performance of our VSTMM, we conducted extensive experiments and comparisons with state-of-the-art methods on three challenging benchmarks, including Charades-STA, ActivityNet Captions, and TACoS. Without bells and whistles, our VSTMM achieves leading performance with a considerable real-time speed.

Recent studies have revealed the vulnerability of graph convolutional networks (GCNs) to edge-perturbing attacks, such as maliciously inserting or deleting graph edges. However, theoretical proof of such vulnerability remains a big challenge, and effective defense schemes are still open issues. In this article, we first generalize the formulation of edge-perturbing attacks and strictly prove the vulnerability of GCNs to such attacks in node classification tasks. Following this, an anonymous GCN, named AN-GCN, is proposed to defend against edge-perturbing attacks. In particular, we present a node localization theorem to demonstrate how GCNs locate nodes during their training phase. In addition, we design a staggered Gaussian noise-based node position generator and a spectral graph convolution-based discriminator (in detecting the generated node positions). Furthermore, we provide an optimization method for the designed generator and discriminator. It is demonstrated that the AN-GCN is secure against edge-perturbing attacks in node classification tasks, as AN-GCN is developed to classify nodes without the edge information (making it impossible for attackers to perturb edges anymore). Extensive evaluations verify the effectiveness of the general edge-perturbing attack (G-EPA) model in manipulating the classification results of the target nodes. More importantly, the proposed AN-GCN can achieve 82.7% in node classification accuracy without the edge-reading permission, which outperforms the state-of-the-art GCN.

Blockchain has revolutionized a variety of fields by providing decentralization, immutability, transparency, and auditability. This paper designs Blockchained Edge Resource Auction (BERA) for edge computing systems to allocate computing resources to application service providers (ASP) in a secure manner. BERA comprises two key components: Blockchain-based Sealed-Bid Auction (BSBA) and Graph Neural Network (GNN)-based Fraud Detection (GFD). BSBA designs smart contracts to realize sealed-bid auctions overhead blockchain. It incorporates the homomorphic commitment technique to guarantee the transactional privacy of ASPs’ bidding information and performs interval membership zero-knowledge proof to verify the legitimacy of auction results. While the privacy-preserving property of BSBA is desirable, the veiled bidding information tends to breed fraudulent behaviors. Therefore, GFD is further proposed to identify abnormal auction behaviors in BSBA without revealing bidding information of ASPs. GFD converts the blockchain data of BSBA to an auction behavioral graph of ASPs, and uses GNN to discover stealth frauds based on interactive patterns. In addition, we design a subgraph extraction scheme for GFD to improve its scalability. We implement BERA on a private Ethereum blockchain and successfully realize edge resource auctions. We simulate several types of auction frauds and identify them with GFD. The experimental results show that our method outperforms other benchmarks.

This paper addresses the temporal sentence grounding (TSG). Although existing methods have made decent achievements in this task, they not only severely rely on abundant video-query paired data for training, but also easily fail into the dataset distribution bias. To alleviate these limitations, we introduce a novel Equivariant Consistency Regulation Learning (ECRL) framework to learn more discriminative query-related frame-wise representations for each video, in a self-supervised manner. Our motivation comes from that the temporal boundary of the query-guided activity should be consistently predicted under various video-level transformations. Concretely, we first design a series of spatio-temporal augmentations on both foreground and background video segments to generate a set of synthetic video samples. In particular, we devise a self-refine module to enhance the completeness and smoothness of the augmented video. Then, we present a novel self-supervised consistency loss (SSCL) applied on the original and augmented videos to capture their invariant query-related semantic by minimizing the KL-divergence between the sequence similarity of two videos and a prior Gaussian distribution of timestamp distance. At last, a shared grounding head is introduced to predict the transform-equivariant query-guided segment boundaries for both the original and augmented videos. Extensive experiments on three challenging datasets (ActivityNet, TACoS, and Charades-STA) demonstrate both effectiveness and efficiency of our proposed ECRL framework.

In the coming era of Metaverse, Augmented Reality (AR) has become a key enabler of diverse applications including healthcare, education, smart cities, and entertainments. To provide users with interactive and immersive experience, most AR applications require extremely high responsiveness and ultra-low processing latency. Mobile edge computing (MEC) has demonstrated great potentials in meeting such stringent latency requirements and resource demands of AR applications, by implementing AR requests in edge servers within the proximity of users. In this paper, we investigate the reward maximization problem for AR applications with uncertain resource demands in an MEC network, such that the accumulative reward of services provided for AR applications is maximized, while ensuring that the responsiveness of AR applications is enhanced, subject to network resource capacity. To this end, we formulate an exact solution when the problem size is small, otherwise we devise an efficient approximation algorithm with a provable approximation ratio for the problem. We also develop an online learning algorithm with a bounded regret for the dynamic reward maximization problem without the knowledge of future arrivals of AR requests, by adopting the Multi-Armed Bandits (MAB) technique. Considering maximizing the reward may defer the implementations of some urgent yet low-award requests, we propose a fairness-aware online learning algorithm for the dynamic reward maximization problem, through a data rate prediction mechanism that adopts a multi-task and multi-timescale Long Short-Term Memory (MT2-LSTM) method. Finally, we evaluate the performance of the proposed algorithms for AR applications by building a real test bed. Experimental results show that the proposed algorithms outperform existing studies by improving the award by 13%.

Unmanned Aerial Vehicles (UAVs) have gained increasing attention by both academic and industrial communities, due to their flexible deployment and efficient line-of-sight communication. Recently, UAVs equipped with base stations have been envisioned as a key technology to provide 5G network services for mobile users. In this article, we provide timely services on the data streams of mobile users in a UAV-aided Mobile Edge Computing (MEC) network, in which each UAV is equipped with a 5G small-cell base station for communication and data processing. Specifically, we first formulate a flow-time minimization problem by jointly caching services and offloading tasks of mobile users to the UAV-aided MEC with the aim to minimize the flow time, where the flow time of a user request is referred to the time duration from the request issuing time point to its completion point, subject to resource and energy capacity on each UAV. We then propose a spatial-temporal learning optimization framework. We also devise an online algorithm with a competitive ratio for the problem based upon the framework, by leveraging the round-robin scheduling and dual fitting techniques. Finally, we evaluate the performance of the proposed algorithms through experimental simulation. The simulation results demonstrate that the proposed algorithms outperform their comparison counterparts, by reducing the flow time no less than 19% on average.

Given an untrimmed video, temporal sentence localization (TSL) aims to localize a specific segment according to a given sentence query. Though respectable works have made decent achievements in this task, they severely rely on dense video frame annotations, which require a tremendous amount of human effort to collect. In this paper, we target another more practical and challenging setting: one-shot temporal sentence localization (one-shot TSL), which learns to retrieve the query information among the entire video with only one annotated frame. Particularly, we propose an effective and novel tree-structure baseline for one-shot TSL, called Multiple Hypotheses Segment Tree (MHST), to capture the query-aware discriminative frame-wise information under the insufficient annotations. Each video frame is taken as the leaf-node, and the adjacent frames sharing the same visual-linguistic semantics will be merged into the upper non-leaf node for tree building. At last, each root node is an individual segment hypothesis containing the consecutive frames of its leaf-nodes. During the tree construction, we also introduce a pruning strategy to eliminate the interference of query-irrelevant nodes. With our designed self-supervised loss functions, our MHST is able to generate high-quality segment hypotheses for ranking and selection with the query. Experiments on two challenging datasets demonstrate that MHST achieves competitive performance compared to existing methods.

Distantly-Supervised Named Entity Recognition (DS-NER) effectively alleviates the data scarcity problem in NER by automatically generating training samples. Unfortunately, the distant supervision may induce noisy labels, thus undermining the robustness of the learned models and restricting the practical application. To relieve this problem, recent works adopt self-training teacher-student frameworks to gradually refine the training labels and improve the generalization ability of NER models. However, we argue that the performance of the current self-training frameworks for DS-NER is severely underestimated by their plain designs, including both inadequate student learning and coarse-grained teacher updating. Therefore, in this paper, we make the first attempt to alleviate these issues by proposing: (1) adaptive teacher learning comprised of joint training of two teacher-student networks and considering both consistent and inconsistent predictions between two teachers, thus promoting comprehensive student learning. (2) fine-grained student ensemble that updates each fragment of the teacher model with a temporal moving average of the corresponding fragment of the student, which enhances consistent predictions on each model fragment against noise. To verify the effectiveness of our proposed method, we conduct experiments on four DS-NER datasets. The experimental results demonstrate that our method significantly surpasses previous SOTA methods. The code is available at https://github.com/zenhjunpro/ATSEN.

Given an untrimmed video, temporal sentence grounding (TSG) aims to locate a target moment semantically according to a sentence query. Although previous respectable works have made decent success, they only focus on high-level visual features extracted from the consecutive decoded frames and fail to handle the compressed videos for query modelling, suffering from insufficient representation capability and significant computational complexity during training and testing. In this paper, we pose a new setting, compressed-domain TSG, which directly utilizes compressed videos rather than fully-decompressed frames as the visual input. To handle the raw video bit-stream input, we propose a novel Three-branch Compressed-domain Spatial-temporal Fusion (TCSF) framework, which extracts and aggregates three kinds of low-level visual features (I-frame, motion vector and residual features) for effective and efficient grounding. Particularly, instead of encoding the whole decoded frames like previous works, we capture the appearance representation by only learning the I-frame feature to reduce delay or latency. Besides, we explore the motion information not only by learning the motion vector feature, but also by exploring the relations of neighboring frames via the residual feature. In this way, a three-branch spatial-temporal attention layer with an adaptive motion-appearance fusion module is further designed to extract and aggregate both appearance and motion information for the final grounding. Experiments on three challenging datasets shows that our TCSF achieves better performance than other state-of-the-art methods with lower complexity.

Vision Transformers (ViTs), which made a splash in the field of computer vision (CV), have shaken the dominance of convolutional neural networks (CNNs). However, in the process of industrializing ViTs, backdoor attacks have brought severe challenges to security. The success of ViTs benefits from the self-attention mechanism. However, compared with CNNs, we find that this mechanism of capturing global information within patches makes ViTs more sensitive to patch-wise triggers. Under such observations, we delicately design a novel backdoor attack framework for ViTs, dubbed BadViT, which utilizes a universal patch-wise trigger to catch the model's attention from patches beneficial for classification to those with triggers, thereby manipulating the mechanism on which ViTs survive to confuse itself. Furthermore, we propose invisible variants of BadViT to increase the stealth of the attack by limiting the strength of the trigger perturbation. Through a large number of experiments, it is proved that BadViT is an efficient backdoor attack method against ViTs, which is less dependent on the number of poisons, with satisfactory convergence, and is transferable for downstream tasks. Furthermore, the risks inside of ViTs to backdoor attacks are also explored from the perspective of existing advanced defense schemes.

Temporal sentence localization in videos (TSLV) aims to retrieve the most interested segment in an untrimmed video according to a given sentence query. However, almost of existing TSLV approaches suffer from the same limitations: (1) They only focus on either frame-level or object-level visual representation learning and corresponding correlation reasoning, but fail to integrate them both; (2) They neglect to leverage the rich semantic contexts to further benefit the query reasoning. To address these issues, in this paper, we propose a novel Hierarchical Visual- and Semantic-Aware Reasoning Network (HVSARN), which enables both visual- and semantic-aware query reasoning from object-level to frame-level. Specifically, we present a new graph memory mechanism to perform visual-semantic query reasoning: For visual reasoning, we design a visual graph memory to leverage visual information of video; For semantic reasoning, a semantic graph memory is also introduced to explicitly leverage semantic knowledge contained in the classes and attributes of video objects, and perform correlation reasoning in the semantic space. Experiments on three datasets demonstrate that our HVSARN achieves a new state-of-the-art performance.

Temporal sentence grounding (TSG) aims to localize the temporal segment which is semantically aligned with a natural language query in an untrimmed video. Most existing methods extract frame-grained features or object-grained features by 3D ConvNet or detection network under a conventional TSG framework, failing to capture the subtle differences between frames or to model the spatio-temporal behavior of core persons/objects. In this paper, we introduce a new perspective to address the TSG task by tracking pivotal objects and activities to learn more fine-grained spatio-temporal behaviors. Specifically, we propose a novel Temporal Sentence Tracking Network (TSTNet), which contains (A) a Cross-modal Targets Generator to generate multi-modal templates and search space, filtering objects and activities, and (B) a Temporal Sentence Tracker to track multi-modal targets for modeling the targets’ behavior and to predict query-related segment. Extensive experiments and comparisons with state-of-the-arts are conducted on challenging benchmarks: Charades-STA and TACoS. And our TSTNet achieves the leading performance with a considerable real-time speed.

With the maturity of depth sensors, the vulnerability of 3D point cloud models has received increasing attention in various applications such as autonomous driving and robot navigation. Previous 3D adversarial attackers either follow the white-box setting to iteratively update the coordinate perturbations based on gradients, or utilize the output model logits to estimate noisy gradients in the black-box setting. However, these attack methods are hard to be deployed in real-world scenarios since realistic 3D applications will not share any model details to users. Therefore, we explore a more challenging yet practical 3D attack setting, i.e., attacking point clouds with black-box hard labels, in which the attacker can only have access to the prediction label of the input. To tackle this setting, we propose a novel 3D attack method, termed 3D Hard-label attacker (3DHacker), based on the developed decision boundary algorithm to generate adversarial samples solely with the knowledge of class labels. Specifically, to construct the class-aware model decision boundary, 3DHacker first randomly fuses two point clouds of different classes in the spectral domain to craft their intermediate sample with high imperceptibility, then projects it onto the decision boundary via binary search. To restrict the final perturbation size, 3DHacker further introduces an iterative optimization strategy to move the intermediate sample along the decision boundary for generating adversarial point clouds with smallest trivial perturbations. Extensive evaluations show that, even in the challenging hard-label setting, 3DHacker still competitively outperforms existing 3D attacks regarding the attack performance as well as adversary quality.

Deep neural networks often exhibit the overconfidence issue when encountering out-of-distribution (OOD) samples. To address this, leveraging large-scale pre-trained models like CLIP has shown promise. While CLIP has the capability to encode a vast array of interconnected concepts, current OOD detection methods based on it primarily focus on ID categories and a limited set of OOD categories. In this paper, we propose a novel approach that harnesses the power of WordNet to fully exploit the rich knowledge encapsulated within CLIP, resulting in enhanced OOD detection performance. Our methodology involves constructing a word tree that includes both in-distribution (ID) words and a large set of semantically similar OOD words selected from WordNet. By matching a test image with the concepts of the words in the word tree using CLIP, we estimate the probability of the image being classified as either ID or OOD. Furthermore, we introduce a conditional random field model to effectively handle both the parent-child and the sibling-sibling conflicts in the concept matching results. Extensive experiments under various ID/OOD settings demonstrate the effectiveness of our approach and its superiority over state-of-the-art methods.

Unsupervised domain adaptive person re-identification has garnered considerable attention due to its practical significance. Previous research has focused on utilizing the teacher-student framework within the clustering and fine-tuning paradigm to reduce the domain gap between different person re-identification datasets. Building upon these recent advancements, we propose a novel approach to further explore the structural information of the teacher and student networks from multiple perspectives, including adaptive sample updating, global feature arrangement, and local discrepancy learning. By integrating these three components, we introduce the Adaptive Structure Learning Framework (ASL) for unsupervised domain adaptive person re-identification. Compared to the baseline method, our approach achieves a significant improvement of 7.3% in mean Average Precision (mAP) on the Market2MSMT task. Furthermore, our experimental results across three benchmark datasets provide further evidence of the effectiveness of our proposed method.

This paper considers online convex optimization (OCO) with generated i.i.d. stochastic constraints, where the performance is measured by adaptive regret. The stochastic constraints are disclosed at each round to the learner after the decision is made. Different from the previous non-adaptive constrained OCO algorithm which is directly generalized from the static online gradient descent algorithm, we propose the novel Virtual Queue-based Following-the-Leader-History (VFLH) strategy to make the constrained OCO algorithm adaptive. In this framework, the algorithm generalizes experts that deal with the static constrained optimization problem within specified time intervals. Subsequently, it combines the predictions of active experts to produce a final choice and unify the average regret and constraints virtual queue. The algorithm’s performance is evaluated based on two metrics: the bounds of constraint violation and regret. To address the difficulty of proving the constraint violation bound under the adaptive setting, we first employ the multi-objective drift analysis approach to handle the constraints virtual queue. Further analysis of the regret bound and the numerical results also supports the performance of the newly proposed algorithm.

Federated Learning enables collaboratively model training among a number of distributed devices with the coordination of a centralized server, where each device alternatively performs local gradient computation and communication to the server. FL suffers from significant performance degradation due to the excessive communication delay between the server and devices, especially when the network bandwidth of these devices is limited, which is common in edge environments. Existing methods overlap the gradient computation and communication to hide the communication latency to accelerate the FL training. However, the overlapping can also lead to an inevitable gap between the local model in each device and the global model in the server that seriously restricts the convergence rate of learning process. To address this problem, we propose a new overlapping method for FL, AOCC-FL, which aligns the local model with the global model via calibrated compensation such that the communication delay can be hidden without deteriorating the convergence performance. Theoretically, we prove that AOCC-FL admits the same convergence rate as the non-overlapping method. On both simulated and testbed experiments, we show that AOCC-FL achieves a comparable convergence rate relative to the non-overlapping method while outperforming the state-of-the-art overlapping methods.

This paper addresses the challenging task of language-driven moment retrieval. Previous methods are typically trained to localize the target moment corresponding to a single sentence query in a complicated video. However, this specific moment generally delivers richer contents than the query, i.e., the semantics of one query may miss certain object details or actions in the complex foreground-background visual contents. Such information imbalance between two modalities makes it difficult to finely align their representations. To this end, instead of training with a single query, we propose to utilize the diversity and complementarity among different queries corresponding to the same video moment for enriching the textual semantics. Specifically, we develop a Teacher-Student Moment Retrieval (TSMR) framework to fill this cross-modal information gap. A teacher model is trained to not only encode a certain query but also capture extra complementary queries to aggregate contextual semantics for obtaining more comprehensive moment-related query representations. Since the additional queries are inaccessible during inference, we further introduce an adaptive knowledge distillation mechanism to train a student model with a single query input by selectively absorbing the knowledge from the teacher model. In this manner, the student model is more robust to the cross-modal information gap during the moment retrieval guided by a single query. Experimental results on two benchmarks demonstrate the effectiveness of our proposed method.

With increasingly strict data privacy regulations, federated learning (FL) has become one of the most often heard machine learning techniques due to its privacy-preserving trait. To efficiently implement the FL intelligence, researchers recently resort to a newly emerged computing paradigm, mobile-edge computing (MEC), and bring about a burst of works. However, most existing works neglect practical issues in MEC systems, e.g., device heterogeneity, unstable channel conditions, and unknown user mobility. Any of them, if not handled properly, can cause fatal failures to FL. This article proposed a novel FL framework, called multicore FL (MC-FL), to help FL intelligence land successfully on realistic MEC systems. A distinct feature of MC-FL is maintaining and training multiple global models (GMs) that exhibit different tradeoffs between learning performances and computational complexity. While this modification seems simple, it can effectively handle the device heterogeneity and device status variations, and improve the compatibility and robustness of FL. Furthermore, MC-FL employs a partial client participation scheme that allows participating clients to vary across time. This enables MC-FL to function under uncertain mobile environments. We rigorously prove the convergence of the designed MC-FL framework. In particular, we propose an online client scheduling scheme for MC-FL to judiciously schedule clients for training multiple GMs in a manner that minimizes the completion time of MC-FL. We also provide a service provisioning scenario with MC-FL to show how service subscribers could benefit from multiple GMs and improve their Quality of Experience (QoE). We evaluate our method on real-world data sets, and the results show that MC-FL outperforms state-of-the-art benchmarks.

Mobile-edge computing (MEC) is a new computing paradigm that enables cloud computing and information technology (IT) services to be delivered at the network’s edge. By shifting the load of cloud computing to individual local servers, MEC helps meet the requirements of ultralow latency, localized data processing, and extends the potential of the Internet of Things (IoT) for end-users. However, the crosscutting nature of MEC and the multidisciplinary components necessary for its deployment have presented additional security and privacy concerns. Fortunately, artificial intelligence (AI) algorithms can cope with excessively unpredictable and complex data, which offers a distinct advantage in dealing with sophisticated and developing adversaries in the security industry. Hence, in this article, we comprehensively provide a survey of security and privacy in MEC from the perspective of AI. On the one hand, we use European Telecommunications Standards Institute (ETSI) MEC reference architecture as our-based framework while merging the software-defined network (SDN) and network function virtualization (NFV) to better illustrate a serviceable platform of MEC. On the other hand, we focus on new security and privacy issues, as well as potential solutions from the viewpoints of AI. Finally, we comprehensively discuss the opportunities and challenges associated with applying AI to MEC security and privacy as possible future research directions.

With the rapid growth of Internet of Things (IoT), smart home develops rapidly in these years, which could assist people who need family medical support. It could integrate health care with ambient assisted living (AAL) technologies and provide activities of daily life (ADLs) to the people who need care. This paper proposes a smart home and cross-cloud-and-edge computing based nursing system (NS). In general, a good NS requires low latency, high stability, and the real-time analysis and response, where the conventional centralized cloud computing based approaches cannot meet those requirements very well. To this end, we introduce a novel distributed joint edge-cloud structure to better satisfy these requirements. Moreover, to deal with the security and privacy issues, we introduce the blockchain to verify the identity of data exchanging and differential-privacy (DP) in the NS to protect the healthcare takers’ data privacy. In a word, we propose a privacy-preserving context-aware multi-armed bandit based online learning approach for edge-cloud-enabled NS via blockchain in IoTs. Additionally, our system with a novel top-down expanding tree based structure can support dynamically increasing health care datasets. Extensive experimental and numerical results demonstrate our solution can achieve accurate recommendation results with sublinear regret performance.

Serverless computing is emerging as an enabling technology for elastic and low-cost AI applications in the edge of core networks. It allows AI developers to decompose a complex training and time-sensitive inference task into multiple functions with dependency, and upload the task to a Multi-access Edge Computing platform (MEC) for execution. Serverless computing adopts a popular design principle: the disaggregation of storage and computation, making the functions ‘stateless’. However, most AI applications are ‘stateful’ and rely on an external storage service to manage their states (ephemeral data). This will incur a prohibitively long delay for delay-sensitive AI applications if external services storing the states are far from the serverless functions. Motivated by this critical issue, in this paper we investigate a fundamental problem in serverless computing – the stateful serverless application placement problem, for which, we first propose an efficient heuristic algorithm, and then devise an approximation algorithm with a provable approximation ratio for one of its special cases. We also consider the online version of the problem, and develop an online learning-driven algorithm with a bounded regret. The crux of the online algorithm is the adoption of the multi-armed bandits technique for dynamic admissions of inference requests, under the uncertainty of both data volumes of requests and network delays. We finally evaluate the performance of the proposed algorithms through experimental simulations. Simulation results show that the proposed algorithms outperform their counterparts, reducing at least 32% in the total cost and 27% of the average delay.

Currently, a high demand for on-device deep neural network (DNN) model deployment is limited by the large model size, computing-intensive floating-point operations (FLOPS), and intellectual property (IP) infringements (i.e., easy access to model duplication for the avoidance of license payments). One appealing solution to addressing the first two concerns is model quantization, which reduces the model size and uses integer operations commonly supported by microcontrollers (MCUs usually do not support FLOPS). To this end, a 1-bit quantized DNN model or deep binary neural network (BNN) significantly improves the memory efficiency, where each parameter in a BNN model has only 1 bit. However, BNN cannot directly provide IP protection (in particular, the functionality of the model is locked unless there is a license payment). In this article, we propose a reconfigurable BNN (RBNN) to further amplify the memory efficiency for resource-constrained Internet of Things (IoT) devices while naturally protecting the model IP. Generally, RBNN can be reconfigured on demand to achieve any one of$M$($M>1$) distinct tasks with the same parameter set, thus only a single task determines the memory requirements. In other words, the memory utilization is improved by a factor of$M$. Our extensive experiments corroborate that up to seven commonly used tasks ($M=7$, six of these tasks are image related and the last one is audio) can co-exist (the value of$M$can be larger). These tasks with a varying number of classes have no or negligible accuracy drop-off (i.e., within 1%) on three binarized popular DNN architectures, including VGG, ResNet, and ReActNet. The tasks span across different domains, e.g., computer vision and audio domains validated herein, with the prerequisite that the model architecture can serve those cross-domain tasks. To fulfill the IP protection of an RBNN model, the reconfiguration can be controlled by both a user key and a device-unique root key generated by the intrinsic hardware fingerprint (e.g., SRAM memory power-up pattern). By doing so, an RBNN model can only be used per paid user per authorized device, thus benefiting both the user and the model provider. The source code is released athttps://github.com/LearningMaker/RBNN.

The volume of data is increasing rapidly, which poses a great challenge for resource-constrained users to process and analyze. A promising approach for solving computation-intensive tasks over big data is to outsource them to the cloud to take advantage of the cloud’s powerful computing capability. However, it also brings privacy and security issues since the data uploaded to the cloud may contain sensitive and private information which should be protected. In this article, we address this problem and focus on the privacy-preserving and secure outsourcing of large-scale nonlinear programming problems (NLPs) subject to both linear constraints and nonlinear constraints. Large-scale NLPs play an important role in the field of data analytics but have not received enough attention in the context of cloud computing. In our outsourcing protocol, we first apply a secure and efficient transformation scheme at the client side to encrypt the private information of the considered NLP. Then, we use the reduced gradient method and generalized gradient method at the server side to solve the transformed large-scale NLPs under linear constraints and nonlinear constraints, respectively. We provide security analysis of the proposed protocol, and evaluate its performance via a series of experiments. The experimental results show that our protocol can efficiently solve large-scale NLPs and save much time for the client, providing a great potential for real applications.

Social network events are related to the national economy and people’s livelihood, so timely perception and processing of massive social network events data are becoming increasingly important for public opinion analysis. How to fully exploit the access feature of event information to store and manage social network events is of great significance for accurate and real-time query analysis. We propose a social network event storage management method based on microblog text clustering and hot/cold data classification. First, for the microblog text data, we construct a keyword provenance graph by using the information entropy to measure the weight of the edge between keyword nodes. Then, we cluster the events using the provenance-based community partition (PCP) with local modularity to improve the event clustering accuracy. In addition, we can further filter noisy data via incremental clustering, enable hot/cold event data classification and dynamic migration, and compress cold data to save space on a hybrid storage architecture. The experimental results show that the clustering purity can reach more than 93% and the query time can be reduced by more than 70% using clustering and hybrid storage policy.

Temporal sentence grounding (TSG) is an important yet challenging task in video-based information retrieval. Given an untrimmed video input, it requires the machine to predict the interested video segment semantically related to a given sentence query. Most existing TSG methods train well-designed deep networks to align the semantic between video-query pairs for activity grounding with a large amount of data. However, we argue that these works easily capture the selection biases of video-query pairs in a dataset rather than showing the robust reasoning abilities to handle the rarely appeared pairs (i.e., few-shot contents). To alleviate such limitation of the off-balance data distribution during the network training, in this paper, we propose a novel memory-augmented network called Memory-Guided Semantic Learning Network (MGSL-Net) to handle the few-shot TSG task for enhancing the model generalization ability. Specifically, given the matched video-query input, we first employ a graph attentive cross-modal interaction module to align their semantics in a cycle-consistent manner. Then, we develop the memory modules in both video and query domains to record the cross-modal shared semantic features in the domain-specific persistent memory. At last, a heterogeneous attention module is utilized to integrate the memory-enhanced multi-modal features in both video and query domains with further feature calibration. During training, the memory modules are dynamically associated with both common and rare cases to memorize all appeared contents, alleviating the issue of forgetting the few-shot contents. Therefore, in testing, the rare cases can be enhanced by retrieving the stored memories, improving the generalization ability of the model. Experimental results on three benchmarks (ActivityNet Caption, Charades-STA and TACoS) show the superiority of our method on both effectiveness and efficiency.

Identifying intrusion from massive and multi-source logs accurately and in real-time presents challenges for today's users. This article presents Paradise, a real-time, generalized, and distributed provenance-based intrusion detection method. Paradise introduces a novel extract strategy to prune and extract process feature vectors from provenance dependencies at the system log level, and it stores them in high-efficiency memory databases. Using this strategy, Paradise does not depend on the specific operating system type or provenance collection framework. Provenance-based dependencies are calculated independently during the detection phase, thus, Paradise can negotiate all detection results from multiple detectors without extra communication overhead between detectors. Paradise also employs an efficient load-balanced distribution scheme that enhances the Kafka architecture to efficiently distribute provenance graph feature vectors to the detectors. The experimental results demonstrate that our method has a high detection accuracy with a low time overhead.

Supporting efficient and secure location-based skyline queries on encrypted data, such as private data outsourced to cloud-based systems, remains an ongoing challenge for efficiency due to significant computational costs in the ciphertext domain. To accelerate privacy-preserving skyline queries, the secure index intuitively contributes to an increase in efficiency. However, designing such a secure index is a challenge while protecting the unlinkability of queries. Meanwhile, there exist little work that can commendably assure efficiency and security. In this paper, we demonstrate SecSky, an efficient solution for supporting secure location-based skyline queries through the secure index. To support SecSky, we devise a novel unified structure, named secure R-tree (SR-tree) index, without privacy leakage (especially indirect privacy). Subsequently, we propose a novel secure location-based dominance protocol, which is utilized to calculate the dominance relationship on the SR-tree. Using this protocol as the building block, our secure location-based skyline query protocol integrates SR-tree, permutation and perturbation techniques to facilitate query processing so as to dramatically reduce the computational overhead. Meanwhile, our proposed solution avoids compromising the privacy of datasets, queries, dominance relationship and skyline results. Finally, we analyze the complexity and security of SecSky. Findings from the experimental evaluation show that our proposed scheme outperforms several other protocols by at least 3 orders of magnitude in terms of query efficiency.

Mobile edge computing (MEC) is envisioned as a promising solution to real-time services in Internet of Vehicles (IoV) by enabling edge caching, computing and communication. However, it is still challenging to implement multimedia streaming in MEC-based IoV due to dynamic vehicular environments and heterogeneous network resources. In this paper, we present an MEC-based architecture for adaptive-bitrate-based (ABR) multimedia streaming in IoV, where each multimedia file is segmented into multiple chunks encoded with different bitrate levels. Then, we formulate a joint resource optimization (JRO) problem by synthesizing heterogeneous edge cache and communication resource constraints, which aims at achieving both smooth play and high-quality service by optimizing chunk placement and transmission. For chunk placement, a multi-armed bandit (MAB) algorithm is proposed for online scheduling with low overhead but slow convergence. Further, a deep-Q-learning algorithm is proposed to improve cache reward and speed up convergence by using replay memory for repeatedly training. For chunk transmission, we design an adaptive-quality-based chunk selection (AQCS) algorithm, which determines bandwidth allocation and quality level based on a benefit function incorporating quality level, available playback time, and freezing delay. Lastly, we build the simulation model and give comprehensive performance evaluation, which demonstrates the superiority of proposed algorithms.

In the traditional video streaming service provisioning paradigm, viewers typically request video content through a central Content Delivery Network (CDN) server. However, because of the uncertain wide area network delays, the (remote) viewers usually suffer from long video streaming delay, which affects the quality of experience. Multi-Access Edge Computing (MEC) offers a way to shorten the video streaming delay by building small-scale cloud infrastructures at the network edge, which are in close proximity to the viewers. In this paper, we present novel centralized and distributed algorithms for the video content placement problem in MEC. In the proposed centralized video content placement algorithm, we leverage the Lyapunov optimization technique to formulate the video content placement problem as a series of one-time-slot optimization problems and apply an Alternating Direction Method of Multipliers (ADMM)-based method to solve each of them. We further devise a distributed Multi-Agent Reinforcement Learning (MARL)-based method with value decomposition mechanism and parallelization policy update method to solve the video content placement problem. The value Decomposition mechanism deals with the credit assignment among multiple agents, which promotes the cooperative optimization of the global target and reduces the frequency of information exchange. The parallelization of policy network can speed up the convergence process. Simulation results verify the effectiveness and superiority of our proposed centralized and distributed algorithms in terms of performance.

Federated learning (FL) has reshaped the learning paradigm by overcoming privacy concerns and siloed data. In FL, an aggregator schedules a set of mobile users (MUs) to collectively train a global model with their local datasets and subsequently aggregates their model updates. However, the users have many uncertainties like unstable network connections and volatile availability, which leads to the straggler problem and deteriorates the efficiency of the FL system. Besides, the issue of non-IID datasets hinders the convergence performance of the global model. To hurdle the user uncertainties, we associate a deadline with the decision in each round and partially collect MUs' updates after the deadline, which can be achieved by considering surplus budget constraints. Moreover, we introduce fairness constraints for the non-IID issue. We propose a deadline-aware task replication for surplus client scheduling policy, called FEDDATE-CS. FEDDATE-CS is developed based on a novel contextual-combinatorial multi-armed bandit (CCMAB) learning framework with fairness guarantee. We extend the hypercube-based CCMAB framework by integrating the Lyapunov queuing technique and rigorously prove that FEDDATE-CS achieves a sublinear regret bound and provides an$[\mathcal{O}(1/V),\mathcal{O}(V)]$regret-fairness tradeoff for any fairness control factor$V>0$. We conduct extensive evaluations to verify the significant superiority of FEDDATE-CS over benchmarks.

Multi-access edge computing (MEC) enables extreme low-latency AI services, such as Augmented Reality (AR) and Virtual Reality (VR), by deploying cloudlets in locations close to users. Meanwhile, a 5G hierarchical service market is emerging with both large-scale and small-scale network service providers competing for both computing and network bandwidth resources of an infrastructure provider. In this paper, we investigate the problem of caching services originally deployed in remote clouds to cloudlets in an MEC network in a hierarchical service market. For the service caching problem, we first propose a novel approximation-restricted framework that guarantees the stability of the 5G service market. Under the proposed framework, we first propose an approximation algorithm with a provable approximation ratio for the problem with non-selfish network service providers. We then design an efficient Stackelberg congestion game with selfish network service providers, and analyze the Price of Anarchy (PoA) of the proposed Stackelberg congestion game to measure the efficiency loss of the game due to selfishness of network service providers. Considering that the request rate of each service may not be given in advance, we study the service caching problem with the uncertainlity of request rates, and propose an approximation algorithm and a Stackelberg game via leveraging the randomized rounding technique. We finally evaluate the performance of the proposed algorithms and mechanisms by both simulations and implementations in a real test-bed. Results show that the performance of our proposed mechanisms achieve around 9.2% less cost than those of existing approaches.

With the development of 5G technology, mobile edge computing is emerging as an enabling technique to reduce the response latency of network services by deploying cloudlets at 5G base stations to form mobile edge cloud (MEC) networks. Network service providers now shift their services from remote clouds to cloudlets of MEC networks in the proximity of users. However, the permanent placement of network services into an MEC network is not economic due to limited computing and bandwidth resources imposed on its cloudlets. A smart way is to cache frequently demanded services from remote clouds to cloudlets of the MEC network. In this paper, we study the problem of service caching in an MEC network under a service market with multiple network service providers competing for both computation and bandwidth resources in terms of Virtual Machines (VMs) in the MEC network. We first propose an Integer Linear Program (ILP) solution and a randomized rounding algorithm, for the problem without VM sharing among different network service providers. We then devise a distributed and stable game-theoretical mechanism for the problem with VM sharing among network service providers, with the aim to minimize the social cost of all network service providers, through introducing a novel cost sharing model and a coalition formation game. We also analyze the performance guarantee of the proposed mechanism, Strong Price of Anarchy (SPoA). We third consider the cost- and delay-sensitive service caching problem with temporal VM sharing, and propose a mechanism with provable SPoA. We finally evaluate the performance through extensive simulations and a real world test-bed implementation. Experimental results demonstrate that the proposed algorithms outperform existing approaches by achieving at least$40\%$lower social cost via service caching and resource sharing among different network service providers.

As an increasingly popular task in multimedia information retrieval, video moment retrieval (VMR) aims to localize the target moment from an untrimmed video according to a given language query. Most previous methods depend heavily on numerous manual annotations (i.e., moment boundaries), which are extremely expensive to acquire in practice. In addition, due to the domain gap between different datasets, directly applying these pre-trained models to an unseen domain leads to a significant performance drop. In this paper, we focus on a novel task: cross-domain VMR, where fully-annotated datasets are available in one domain (“source domain”), but the domain of interest (“target domain”) only contains unannotated datasets. As far as we know, we present the first study on cross-domain VMR. To address this new task, we propose a novelMulti-ModalCross-DomainAlignment (MMCDA) network to transfer the annotation knowledge from the source domain to the target domain. However, due to the domain discrepancy between the source and target domains and the semantic gap between videos and queries, directly applying trained models to the target domain generally leads to a performance drop. To solve this problem, we develop three novel modules: (i) a domain alignment module is designed to align the feature distributions between different domains of each modality; (ii) a cross-modal alignment module aims to map both video and query features into a joint embedding space and to align the feature distributions between different modalities in the target domain; and (iii) a specific alignment module tries to obtain the fine-grained similarity between a specific frame and the given query for optimal localization. By jointly training these three modules, our MMCDA can learn domain-invariant and semantic-aligned cross-modal representations. Extensive experiments on three challenging benchmarks (ActivityNet Captions, Charades-STA and TACoS) illustrate that our cross-domain method MMCDA outperforms all state-of-the-art single-domain methods. Impressively, MMCDA raises the performance by more than 7% in representative cases, which demonstrates its effectiveness.

Temporal sentence grounding aims to localize a target segment in an untrimmed video semantically according to a given sentence query. Most previous works focus on learning frame-level features of each whole frame11In this paper, the frame is a general concept for an actual video frame or a video clip which consists of a few consecutive frames.in the entire video, and directly match them with the textual information. Such frame-level feature extraction leads to the obstacles of these methods in distinguishing ambiguous video frames with complicated contents and subtle appearance differences, thus limiting their performance. In order to differentiate fine-grained appearance similarities among consecutive frames, some state-of-the-art methods additionally employ a detection model like Faster R-CNN to obtain detailed object-level features in each frame for filtering out the redundant background contents. However, these methods suffer from missing motion analysis since the object detection module in Faster R-CNN lacks temporal modeling. To alleviate the above limitations, in this paper, we propose a novelMotion- andAppearance-guided3DSemanticReasoningNetwork (MA3SRN), which incorporates optical-flow-guided motion-aware, detection-based appearance-aware, and 3D-aware object-level features to better reason the spatial-temporal object relations for accurately modelling the activity among consecutive frames. Specifically, we first develop three individual branches for motion, appearance, and 3D encoding separately to learn fine-grained motion-guided, appearance-guided, and 3D-aware object features, respectively. Then, both motion and appearance information from corresponding branches are associated to enhance the 3D-aware features for the final precise grounding. Extensive experiments on three challenging datasets (ActivityNet Caption, Charades-STA and TACoS) demonstrate that the proposed MA3SRN model achieves a new state-of-the-art.

Visual relationship understanding plays an indispensable role in grounded language tasks like visual question answering (VQA), which often requires precisely reasoning about relations among objects depicted in the given question. However, prior works generally suffer from the deficiencies as follows, (1)spatial-relation inference ambiguity, it is challenging to accurately estimate the distance of a pair of visual objects in 2D space if there is a visual-overlap between their 2D bounding-boxes, and (2) language-visual relational alignment missing, it is insufficient to generate a high-quality answer to the question if there is a lack of alignment in the language-visual relations of objects during fusion, even using a powerful fusion model like Transformer. To this end, we first model the spatial relation of a pair of objects in 3D space by augmenting the original 2D bounding-box with 1D depth information, and then propose a novel model namedDepth-awareSemanticGuided RelationalAttention Network (DSGANet), to explicitly exploit the formed 3D spatial relations of objects in an intra-/inter-modality manner for precise relational alignment. Extensive experiments conducted on the benchmarks (VQA v2.0 and GQA) demonstrate DSGANet achieves competitive performance compared to pretrained and non-pretrained models, such as 72.7% vs. 74.6% based on the learned grid features on VQA v2.0.

The emergence of containers dramatically simplifies and facilitates the development and deployment of applications. More and more enterprises deploy their applications on the container cloud platform. For cloud service providers, an effective container workload prediction method is a must to achieve efficient utilization of cloud resources. However, the existing methods are either rarely based on container load characteristics or cannot make accurate real-time predictions. In this paper, we propose a Docker container workload proactive prediction method using a hybrid model combining triple exponential smoothing and long short-term memory (LSTM), which not only can capture both short-term and long-term dependencies in container resource time series but also smooth the container resource utilization data. In order to improve the prediction accuracy of the hybrid model, those two single models are combined using the mean absolute percentage error (MAPE) method. Besides, we design a real-time Docker workload prediction system for the hybrid model. Our experiments show that the mean absolute percentage error of the hybrid model is decreased by an average of 3.24%, 12.18%, 13.42%, 43.45%, and 50.69% compared with the LSTM, the triple exponential smoothing, ES-ARIMA, Bayesian Ridge Regression and BiLSTM with an acceptable time and computational cost overhead.

This article targets the task of language-based video moment localization. The language-based setting of this task allows for an open set of target activities, resulting in a large variation of the temporal lengths of video moments. Most existing methods prefer to first sample sufficient candidate moments with various temporal lengths, then match them with the given query to determine the target moment. However, candidate moments generated with a fixed temporal granularity may be suboptimal to handle the large variation in moment lengths. To this end, we propose a novel multi-stage Progressive Localization Network (PLN) that progressively localizes the target moment in a coarse-to-fine manner. Specifically, each stage of PLN has a localization branch and focuses on candidate moments that are generated with a specific temporal granularity. The temporal granularities of candidate moments are different across the stages. Moreover, we devise a conditional feature manipulation module and an upsampling connection to bridge the multiple localization branches. In this fashion, the later stages are able to absorb the previously learned information, thus facilitating the more fine-grained localization. Extensive experiments on three public datasets demonstrate the effectiveness of our proposed PLN for language-based moment localization, especially for localizing short moments in long videos.

This study is motivated by the maximum connected coverage problem (MCCP), which is to deploy a connected UAV network with given$K$UAVs in the top of a disaster area such that the number of users served by the UAVs is maximized. The deployed UAV network must be connected, since the received data by a UAV from its served users need to be sent to the Internet through relays of other UAVs. Motivated by this application, in this paper we study a more generalized problem – the$h$-hop independently submodular maximization problem, where the MCCP problem is one of its special cases with$h=4$. We propose a$\frac {1-1/e}{2h+3}$-approximation algorithm for the$h$-hop independently submodular maximization problem, where$e$is the base of the natural logarithm. Then, one direct result is a$\frac {1-1/e}{11}$-approximate solution to the MCCP problem with$h=4$, which significantly improves its currently best$\frac {1-1/e}{32}$-approximate solution. We finally evaluate the performance of the proposed algorithm for the MCCP problem in the application of deploying UAV networks, and experimental results show that the number of users served by deployed UAVs delivered by the proposed algorithm is up to 12.5% larger than those by existing algorithms.

Federated learning (FL) is a distributed machine learning technique that enables model development on user equipments (UEs) locally, without violating their data privacy requirements. Conventional FL adopts a single parameter server to aggregate local models from UEs, and can suffer from efficiency and reliability issues – especially when multiple users issue concurrentFL requests. Hierarchical FL consisting of a master aggregator and multiple worker aggregators to collectively combine trained local models from UEs is emerging as a solution to efficient and reliable FL. The placement of worker aggregators and assignment of UEs to worker aggregators plays a vital role in minimizing the cost of implementing FL requests in a Mobile Edge Computing (MEC) network. Cost minimization associated with joint worker aggregator placement and UE assignment problem in an MEC network is investigated in this work. An optimization framework for FL and an approximation algorithm with an approximation ratio for a single FL request is proposed. Online worker aggregator placements and UE assignments for dynamic FL request admissions with uncertain neural network models, where FL requests arrive one by one without the knowledge of future arrivals, is also investigated by proposing an online learning algorithm with a bounded regret. The performance of the proposed algorithms is evaluated using both simulations and experiments in a real testbed with its hardware consisting of server edge servers and devices and software built upon an open source hierarchical FedML (HierFedML) environment. Simulation results show that the performance of the proposed algorithms outperform their benchmark counterparts, by reducing the implementation cost by at least 15% per FL request. Experimental results in the testbed demonstrate the performance gain using the proposed algorithms using real datasets for image identification and text recognition applications.

Mobile crowdsensing (MCS) has recently received considerable attention due to its capability of providing a promising paradigm to complete complex sensing tasks. Existing works on MCS mainly focus on designing incentive mechanisms to attract mobile users to participate in crowdsensing, while ignoring the freshness of information, i.e.,Age of Information(AoI). Although multiple source nodes with common observation can indeed improve the data quality of MCS, it complicates the calculation of the AoI. To address this issue, this article proposes a freshness-aware incentive mechanism in MCS, which not only captures the conflict interests/competitions among users, but also considers the age of information (AoI). Specifically, we define two data sampling models, namedsampling-at-willmodel andsampling-predeterminedmodel. For both models, we design efficient auction mechanisms, which recruit appropriate mobile users, determine the payments, and schedule the data sampling, so as to optimize the average AoI and data quality under budget constraint. It is proved that the proposed auction achieves several desirable properties, including individual rationality, budget balance, truthfulness and computational efficiency. We also theoretically derive the upper bound of the average AoI obtained by the proposed scheme. Finally, we conduct simulations to evaluate the efficiency of the proposed mechanism in optimizing the data quality and AoI.

Temporal sentence grounding (TSG) is crucial and fundamental for video understanding. Although existing methods train well-designed deep networks with large amount of data, we find that they can easily forget the rarely appeared cases during training due to the off-balance data distribution, which influences the model generalization and leads to unsatisfactory performance. To tackle this issue, we propose a memory-augmented network, called Memory-Guided Semantic Learning Network (MGSL-Net), that learns and memorizes the rarely appeared content in TSG task. Specifically, our proposed model consists of three main parts: cross-modal interaction module, memory augmentation module, and heterogeneous attention module. We first align the given video-query pair by a cross-modal graph convolutional network, and then utilize memory module to record the cross-modal shared semantic features in the domain-specific persistent memory. During training, the memory slots are dynamically associated with both common and rare cases, alleviating the forgetting issue. In testing, the rare cases can thus be enhanced by retrieving the stored memories, leading to better generalization. At last, the heterogeneous attention module is utilized to integrate the enhanced multi-modal features in both video and query domains. Experimental results on three benchmarks show the superiority of our method on both effectiveness and efficiency, which substantially improves the accuracy not only on the entire dataset but also on the rare cases.

Temporal video grounding (TVG) aims to localize a target segment in a video according to a given sentence query. Though respectable works have made decent achievements in this task, they severely rely on abundant video-query paired data, which is expensive to collect in real-world scenarios. In this paper, we explore whether a video grounding model can be learned without any paired annotations. To the best of our knowledge, this paper is the first work trying to address TVG in an unsupervised setting. Considering there is no paired supervision, we propose a novel Deep Semantic Clustering Network (DSCNet) to leverage all semantic information from the whole query set to compose the possible activity in each video for grounding. Specifically, we first develop a language semantic mining module, which extracts implicit semantic features from the whole query set. Then, these language semantic features serve as the guidance to compose the activity in video via a video-based semantic aggregation module. Finally, we utilize a foreground attention branch to filter out the redundant background activities and refine the grounding results. To validate the effectiveness of our DSCNet, we conduct experiments on both ActivityNet Captions and Charades-STA datasets. The results demonstrate that our DSCNet achieves competitive performance, and even outperforms most weakly-supervised approaches.

This paper addresses temporal sentence grounding. Previous works typically solve this task by learning frame-level video features and align them with the textual information. A major limitation of these works is that they fail to distinguish ambiguous video frames with subtle appearance differences due to frame-level feature extraction. Recently, a few methods adopt Faster R-CNN to extract detailed object features in each frame to differentiate the fine-grained appearance similarities. However, the object-level features extracted by Faster R-CNN suffer from missing motion analysis since the object detection model lacks temporal modeling. To solve this issue, we propose a novel Motion-Appearance Reasoning Network (MARN), which incorporates both motion-aware and appearance-aware object features to better reason object relations for modeling the activity among successive frames. Specifically, we first introduce two individual video encoders to embed the video into corresponding motion-oriented and appearance-aspect object representations. Then, we develop separate motion and appearance branches to learn motion-guided and appearance-guided object relations, respectively. At last, both motion and appearance information from two branches are associated to generate more representative features for final grounding. Extensive experiments on two challenging datasets (Charades-STA and TACoS) show that our proposed MARN significantly outperforms previous state-of-the-art methods by a large margin.

Graph neural networks (GNNs) have achieved state-of-the-art performance in many graph-based tasks such as node classification and graph classification. However, many recent works have demonstrated that an attacker can mislead GNN models by slightly perturbing the graph structure. Existing attacks to GNNs are either under the less practical threat model where the attacker is assumed to access the GNN model parameters, or under the practical black-box threat model but consider perturbing node features that are shown to be not enough effective. In this paper, we aim to bridge this gap and consider black-box attacks to GNNs with structure perturbation as well as with theoretical guarantees. We propose to address this challenge through bandit techniques. Specifically, we formulate our attack as an online optimization with bandit feedback. This original problem is essentially NP-hard due to the fact that perturbing the graph structure is a binary optimization problem. We then propose an online attack based on bandit optimization which is proven to be sublinear to the query number T, i.e., O(✓NT3/4) where N is the number of nodes in the graph. Finally, we evaluate our proposed attack by conducting experiments over multiple datasets and GNN models. The experimental results on various citation graphs and image graphs show that our attack is both effective and efficient.

Spatial-Temporal Video Grounding (STVG) is a challenging task which aims to localize the spatio-temporal tube of the interested object semantically according to a natural language query. Most previous works not only severely rely on the anchor boxes extracted by Faster R-CNN, but also simply regard the video as a series of individual frames, thus lacking their temporal modeling. Instead, in this paper, we are the first to propose an anchor-free framework for STVG, called Gaussian Kernel-based Cross Modal Network (GKCMN). Specifically, we utilize the learned Gaussian Kernel-based heatmaps of each video frame to locate the query-related object. A mixed serial and parallel connection network is further developed to leverage both spatial and temporal relations among frames for better grounding. Experimental results on VidSTG dataset demonstrate the effectiveness of our proposed GKCMN.

This paper addresses the problem of temporal moment localization with natural language, which aims to localize a target video moment according to a language description. A key challenge in this task is how to learn effective alignment between vision and language features extracted from an untrimmed video and a sentence description. Traditional methods generally utilize the vanilla attention mechanism to simulate soft alignment between extracted video and sentence features, however, they lack the training signal to alleviate the domain gap between video and sentence domains. In this paper, we propose a novel cross-domain alignment module to deeply align the features in two domains, yielding to learn better representations. We evaluate our model on three public datasets: Charades-STA, ActivityNet Captions and TACoS, and the experiments prove that our proposed cross-domain alignment module can bring significant improvement, and can be also applied to other baseline models.

This study is motivated by the maximum connected coverage problem (MCCP), which is to deploy a connected UAV network with given K UAVs in the top of a disaster area such that the number of users served by the UAVs is maximized. The deployed UAV network must be connected, since the received data by a UAV from its served users need to be sent to the Internet through relays of other UAVs. Motivated by this application, in this paper we study a more generalized problem – the h-hop independently submodular maximization problem, where the MCCP problem is one of its special cases with h = 4. We propose a $\frac{{1 - 1/e}}{{2h + 3}}$-approximation algorithm for the h-hop independently submodular maximization problem, where e is the base of the natural logarithm. Then, one direct result is a $\frac{{1 - 1/e}}{{11}}$-approximate solution to the MCCP problem with h = 4, which significantly improves its currently best $\frac{{1 - 1/e}}{{32}}$-approximate solution. We finally evaluate the performance of the proposed algorithm for the MCCP problem in the application of deploying UAV networks, and experimental results show that the number of users served by deployed UAVs delivered by the proposed algorithm is up to 12.5% larger than those by existing algorithms.

The age of data (AoD) is identified as one of the most novel and important metrics to measure the quality of big data analytics for Internet-of-Things (IoT) applications. Meanwhile, mobile edge computing (MEC) is envisioned as an enabling technology to minimize the AoD of IoT applications by processing the data in edge servers close to IoT devices. In this paper, we study the AoD minimization problem for IoT big data processing in MEC networks. We first propose an exact solution for the problem by formulating it as an Integer Linear Program (ILP). We then propose an efficient heuristic for the offline AoD minimization problem. We also devise an approximation algorithm with a provable approximation ratio for a special case of the problem, by leveraging the parametric rounding technique. We thirdly develop an online learning algorithm with a bounded regret for the online AoD minimization problem under dynamic arrivals of IoT requests and uncertain network delay assumptions, by adopting the Multi-Armed Bandit (MAB) technique. We finally evaluate the performance of the proposed algorithms by extensive simulations and implementations in a real test-bed. Results show that the proposed algorithms outperform existing approaches by reducing the AoD around 10%.

Crowd counting is a regression task that estimates the number of people in a scene image, which plays a vital role in a range of safety-critical applications, such as video surveillance, traffic monitoring and flow control. In this paper, we investigate the vulnerability of deep learning based crowd counting models to backdoor attacks, a major security threat to deep learning. A backdoor attack implants a backdoor trigger into a target model via data poisoning so as to control the model's predictions at test time. Different from image classification models on which most of existing backdoor attacks have been developed and tested, crowd counting models are regression models that output multi-dimensional density maps, thus requiring different techniques to manipulate. In this paper, we propose two novel Density Manipulation Backdoor Attacks (DMBA- and DMBA+) to attack the model to produce arbitrarily large or small density estimations. Experimental results demonstrate the effectiveness of our DMBA attacks on five classic crowd counting models and four types of datasets. We also provide an in-depth analysis of the unique challenges of backdooring crowd counting models and reveal two key elements of effective attacks: 1) full and dense triggers and 2) manipulation of the ground truth counts or density maps. Our work could help evaluate the vulnerability of crowd counting models to potential backdoor attacks.

X-armed bandits have achieved the state-of-the-art performance in optimizing unknown stochastic continuous functions, which can model many machine learning tasks, specially in big data-driven personalized recommendation. However, bandit algorithms are vulnerable to adversarial attacks. Existing works mainly focus on attacking multi-armed bandits in discrete setting; nevertheless, the attacks against X-armed bandits in continuous setting have not been well explored. In this paper, we aim to bridge this gap and investigate the robustness problem for the X-armed bandits. Specifically, we consider data poisoning attack and propose an attack algorithm named Confidence Poisoning Attack algorithm, which could hijack the clean tree-based X-armed bandits algorithm, i.e., high confidence tree (HCT) and make it choose the nodes including the arm targeted by the attacker very frequently with a sub-linear attack cost, i.e., O(Tα)(0 <α< 1), where T is the total number of rounds. We evaluate the efficiency of our proposed attack algorithm through theoretical analysis and experiments.

As a continuous variant of Multi-armed bandits (MAB), $\mathcal{X}$-armed bandits have enriched many applications of online machine learning like personalized recommendation system. However, the attack and defense to the $\mathcal{X}$-armed bandits remain largely unexplored, though the MAB has proved to be vulnerable. In this paper, we aim to bridge this gap and investigate the robustness analysis for the $\mathcal{X}$-armed bandits. Specifically, we consider action-manipulation attack, which is practical but harder than the existing reward-manipulation attack. We propose an attack algorithm based on a lower bound tree (LBT), which can continuously hijack the learner’s action by perturbing $\mathcal{X}$-armed bandits’ high confidence tree (HCT) construction. As a result, the nodes including the arm targeted by the attacker is selected frequently with a sublinear attack cost. To defend against the LBT attack, we propose a robust version of the HCT algorithm, called RoHCT. We theoretically analyze that the regret of RoHCT is related to the upper bound of the total cost Q and still sublinear to total number of rounds T. We carry out experiments to evaluate the effectiveness of LBT and RoHCT.

Vehicular ad hoc network (VANET) has become an accessible technology for improving road safety and driving experience, the problems of heterogeneity and lack of resources it faces have also attracted widespread attention. With the development of software-defined networking (SDN) and multiaccess edge computing (MEC), a variety of resource allocation strategies in MEC-enabled software-defined networking-based VANET (SDVN) have been proposed to solve these problems. However, we note that few of these work involves the situation where SDVN is under Distributed Denial of Service (DDoS) attacks. Actually, Internet of Things (IoT) devices are extremely easy to be compromised by malicious users, and compromised IoT devices may be used to launch edge DDoS attacks against the MEC servers in MEC-enabled SDVN at any time. In this article, we propose a graph neural network (GNN)-based collaborative deep reinforcement learning (GCDRL) model to generate the resource provisioning and mitigating strategy. The model evaluates the trust value of the vehicles, formulates mitigation of edge DDoS attacks and resource provisioning strategies to ensure that the MEC servers can work normally under edge DDoS attacks. In addition, GNN is adopted in the DRL model to extract the structure feature of the graph composed of MEC servers, and help transfer computing tasks between MEC servers to alleviate the problem of resources imbalance between them. Experimental results show that the method of estimating the vehicular trust value is effective, and our method can make the average throughput of edge nodes more stable and lower down the average delay and the average energy consumption under the edge DDoS attack. Also, a real-world case study is conducted to verify our conclusion.

Age of Information (AoI) has become a new performance metric that quantifies the freshness of information in the Internet of Things (IoT). To optimize the AoI, the latest information should be frequently sampled and timely updated by source nodes (SNs), which, however, contradicts with the fact that the resources of both the SNs and destination node are limited. To consider this issue, this article formulates a more general multisource information update problem, taking into account both the budget constraint of destination node and the limited sampling/updating capabilities of the SNs. Besides that, two different sampling models, namedsampling-predeterminedandsampling-at-willmodels, have been studied, respectively. Particularly, for the information update problem under the sampling-predetermined model, we prove that it is an NP-hard problem. Then, we propose a greedy algorithm to select the appropriate SNs, and theoretically analyze the bound of the AoI achieved by the proposed algorithm. For the sampling-at-will model, we theoretically derive the minimum AoI that can be achieved under given number of updates, based on which, we design an optimal SNs selection and updating time determination mechanism, to achieve the minimum AoI. Finally, we conduct simulations to evaluate the effectiveness of the proposed algorithms.

Federated Learning has attracted widespread attention in recent years because it allows massive edge nodes to collaboratively train machine learning models without sharing their private data sets. However, these edge nodes are usually heterogeneous in computational capability and statistically different in data distribution, i.e., non-independent and identically distributed (IID), leading to significant performance degradation. Although status quo asynchronous training methods can solve the heterogeneity issue, they cannot prevent the non-IID problem from reducing the convergence rate. In this article, we propose a novel paradigm that schedules the gradient with partially averaged gradients and applies the global momentum (GSGM) for asynchronous training over non-IID data sets in an edge environment. Our key idea is to apply global momentum and partial average on the biased gradients calculated on edge nodes after scheduling, to make the training process stable. Empirical results demonstrate that GSGM can well adapt to different degrees of non-IID data and bring 20% performance gains in terms of training stability for popular optimization algorithms with enhanced accuracy over Fashion-Mnist and CIFAR-10 data sets.

Smart home Internet of Things (IoT) has a vibrant market with a wide range of appliances and sensors, spanning across smart home, smart city, and smart factory. However, the security and privacy of these IoT systems have raised serious concerns. Currently, most IoT devices rely on electromagnetic wave-based radio frequency (RF) for communication. Yet, RF has several inherent limitations, such as shortage of spectrum, susceptible to interference, and vulnerable to eavesdropping or jamming attacks. This article presents URadio, a wideband ultrasonic communication system. By leveraging recent advances in reduced Graphene Oxide (rGO), we design a new type of electrostatic ultrasonic transducer, which can achieve more than$6\times $bandwidth than commercial ultrasonic transducers. With this new transducer, we design an OFDM communication system to maximize its data rate for smart home applications. We build a prototype of URadio on a wireless testbed and evaluate its performance in several real-world environments. Our experiments show that URadio can reach up to 360 kb/s data rate at a distance of 81 cm or 20 Kb/s data rate at a distance of 20 m, which supports a variety of smart home applications. We further showcase URadio’s resilience against eavesdropping and jamming attacks, as well as demonstrate its capability of securely localizing objects in an indoor environment.

To facilitate the emerging applications in 5G networks, mobile network operators will provide many network functions in terms of control and prediction. Recently, they have recognized the power of machine learning (ML) and started to explore its potential to facilitate those network functions. Nevertheless, the current ML models for network functions are often derived in an offline manner, which is inefficient due to the excessive overhead for transmitting a huge volume of dataset to remote ML training clouds and failing to provide the incremental learning capability for the continuous model updating. As an alternative solution, we proposeCocktail, an incremental learning framework within a reference 5G network architecture. To achieve cost efficiency while increasing trained model accuracy, an efficient online data scheduling policy is essential. To this end, we formulate an online data scheduling problem to optimize the framework cost while alleviating the data skew issue caused by the capacity heterogeneity of training workers from the long-term perspective. We exploit the stochastic gradient descent to devise an online asymptotically optimal algorithm, including two optimal policies based on novel graph constructions for skew-aware data collection and data training. Small-scale testbed and large-scale simulations validate the superior performance of our proposed framework.

Supporting secure location-based services on encrypted data that is outsourced to cloud computing platforms remains an ongoing challenge for efficiency due to expensive ciphertext calculation overhead. Furthermore, since the clouds may not be trustworthy or even malicious, data security and result authenticity has caused huge concerns. Unfortunately, little work can enable query efficiency, dataset confidentiality and result authenticity to be commendably guaranteed. In this paper, we demonstrate the potential of supporting secure and verifiable location-based skyline queries (SVLSQ). First, we devise a novel and unified structure, named semi-blind R-tree (SR-tree), which protects the query unlinkability. Based on SR-tree, we propose an authenticated data structure, named secure and verifiable scope R-tree (SVSR-tree). Then, we develop several secure protocols based on SVSR-tree to accelerate the query efficiency and reduce the size of verification objects. Our method avoids compromising the privacy of datasets, queries, results and access patterns. Meanwhile, it authenticates the soundness and completeness of the skyline results while preserving privacy. Finally, we analyze the complexity and security of SVLSQ. Findings from the performance evaluation illustrate that SVLSQ is a dramatically efficient method in terms of query (no less than 3 orders of magnitude faster than other solutions) and verification.

Mobile edge computing (MEC) is an enabling technology for low-latency AI applications, by caching AI services originally deployed in remote data centers to 5G base stations in network edge. Due to limited computing resource of 5G base stations, not all services can be cached in base stations to meet the resource demands of user requests. Also, if the workload of a 5G base station reaches to its resource capacity, the energy consumption of the base station will be pushed up exponentially. To reduce the energy consumption and overcome resource limitations on base stations, an alternative is to allow the base stations to collaborate with each other to admit user requests. In this paper, we investigate the problem of collaborative service caching and request offloading between a 5G-enabled MEC and remote data centers, while meeting the quality of service (QoS) requirements of users, and resource capacities on base stations that are operated by multiple selfish network service providers. We aim to maximize the total payoff of all base stations. To this end, we first propose a two-stage optimization framework: In the first stage, we develop a mechanism that adopts a best-reply rule for dynamically distributed coalition formation. In the second stage, we propose a near-optimal payoff allocation method by devising a randomized algorithm with a provable approximation ratio. We then evaluate the performance of the proposed optimization framework by extensive experimental simulations. Simulation results show that the proposed framework outperforms its counterparts by achieving at least 30% higher payoff and 20% lower energy consumption of base stations.

With the development of data plane programmable Software-Defined Networking (SDN), Distributed Denial of Service (DDoS) attacks on the data plane increasingly become fatal. Currently, traditional attack detection methods are mainly used to detect whether a DDoS attack occurs and it is difficult to find the path that the attack flow traverses the network, which makes it difficult to accurately mitigate DDoS attacks. In this article, we propose a detection method based on Spatial-Temporal Graph Convolutional Network (ST-GCN) over the data plane programmable SDN, which maps the network into a graph. It senses the state of switches through In-band Network Telemetry (INT) with sampling, inputs the network state into the spatial-temporal graph convolutional network detection model, and finally finds out the switches through which DDoS attack flows pass. Based on this, we propose a defense method combined with an enhanced whitelist and a precise dropping strategy, which can effectively mitigate DDoS attacks and minimize the impact on legitimate network traffic. The evaluation results show that our detection method can accurately detect the path that the DDoS attack flows pass through, and can effectively mitigate the DDoS attack. Compared to classic methods, our method improves the detection accuracy by nearly 10%. At the same time, the southbound interface load and CPU overhead brought by our detection and defense process are much lower than the classic methods.

The Internet of Things (IoT) has become ubiquitous and greatly affected peoples’ daily lives. With the increasing development of IoT devices, the corresponding security issues are becoming more and more challenging. Such a severe security situation raises the following questions that need urgent attention: What are the primary security threats that IoT devices face currently? How do vendors and users deal with these threats? In this article, we aim to answer these critical questions through a large-scale systematic study. Specifically, we perform a ten-month-long empirical study on the vulnerability of 1,362,906 IoT devices varying from six types. The results show sufficient evidence that N-days vulnerability is seriously endangering the IoT devices: 385,060 (28.25 percent) devices suffer from at least one N-days vulnerability. Moreover, 2669 of these vulnerable devices may have been compromised by botnets. We further reveal the massive differences among five popular IoT search engines:Shodan[1],Censys[2], [3],Zoomeye[4],Fofa[5], andNTI[6]. To study whether vendors and users adopt defenses against the threats, we measure the security of MQTT [7] servers, and identify that 12740 (88 percent) MQTT servers have no password protection. Our analysis can serve as an important guideline for investigating the security of IoT devices, as well as advancing the development of a more secure environment for IoT systems.

Triangle count is a critical parameter in mining relationships among people in social networks. However, directly publishing the findings obtained from triangle counts may bring potential privacy concern, which raises great challenges and opportunities for privacy-preserving triangle counting. In this paper, we choose to use differential privacy to protect triangle counting for large scale graphs. To reduce the large sensitivity caused in large graphs, we propose a novel graph projection method that can be used to obtain an upper bound for sensitivity in different distributions. In particular, we publish the triangle counts satisfying the node-differential privacy with two kinds of histograms: the triangle count distribution and the cumulative distribution. Moreover, we extend the research on privacy preserving triangle counting to one of its applications, the local clustering coefficient. Experimental results show that the cumulative distribution can fit the real statistical information better, and our proposed mechanism has achieved better accuracy for triangle counts while maintaining the requirement of differential privacy.

Large volumes of video data recorded by the increasing mobile devices and embedded sensors can be leveraged to answer queries of our lives, physical world and our evolving society. Especially, the rapid development of convolutional neural networks (CNNs) in the past few years offers the great advantage for multiple tasks in video analysis. However, adopting running CNNs directly on mobile devices and embedded sensors for video analytics brings heavy burden due to their limited capacity, especially for learning a large volume of data. A promising approach is to outsource the computation-intensive part of CNN to cloud. However, the reveal of data to cloud may cause privacy leakage. In addition, the cloud-assisted approach may also bring some communication efficiency challenges for large volume of data. To address both privacy and efficiency issues, we design a privacy-preserving and computationally efficient framework for mobile video analytics. To protect the private information, we split the CNN model into two subnetworks, and first part is used as a feature extractor deployed in the mobile side and the second part is utilized as a classifier deployed in the cloud side. A specific-designed adversarial training process is adopted in order to extract features for normal task classification while hiding the features for sensitive task. In addition, to improve video process efficiency, we design a two-stage framework. The first stage is to extract key frames and necessary intermediate frames, while skipping redundant ones. The second stage is to extract the features of key frames by CNN-based feature extractor but apply optical-flow-based feature propagation algorithm to obtain the features of intermediate frames. Extensive experiments demonstrate our proposed system PrivacyEye can effectively protect private information while keep the accuracy of the normal tasks with less than 2 percent drop, and it saves up to 82.9 percent execution time and 78.8 percent energy consumption.

In this article, a novel edge computing system is proposed for image recognition via memristor-based blaze block circuit, which includes a memristive convolutional neural network (MCNN) layer, two single-memristive blaze blocks (SMBBs), four double-memristive blaze blocks (DMBBs), a global Avg-pooling (GAP) layer, and a memristive full connected (MFC) layer. SMBBs and DMBBs mainly utilize the depthwise separable convolution neural network (DwCNN) that can be implemented with a much smaller memristor crossbar (MC). In the backward propagation, we use batch normalization (BN) layers to accelerate the convergence. In the forward propagation, this circuit combines DwCNN layers/CNN layers with nonseparate BN layers, which means that the required number of operational amplifiers is cut by half as long as the greatly reduced power consumption. A diode is added after the rectified linear unit (ReLU) layer to limit the output of the circuit below the threshold voltage$V_{t}$of the memristor; thus, the circuit is more stable. Experiments show that the proposed memristor-based circuit achieves an accuracy of 84.38% on the CIFAR-10 data set with advantages in computing resources, calculation time, and power consumption. Experiments also show that, when the number of multistate conductance is 28and the quantization bit of the data is 8, the circuit can achieve its best balance between power consumption and production cost.

Caching popular contents in close proximity to the users can effectively reduce the latency in communication systems. Considering the imperfect content fetching from the original server, the caching management at the edge node should be determined according to not only the content popularity but also how difficult to fetch these content objects from their corresponding original servers. In this paper, we study the edge caching for minimizing the long-term average delay, where the fetching delay of the uncached contents is introduced. We construct a decision-theoretic framework for this delay-optimal content-caching problem, where the main obstacle is that the consideration of the imperfect content fetching breaks the Markovian property in the decision-theoretic framework. To overcome this obstacle, by analyzing the queue dynamics within the content fetching delay after the content object is removed from the cache, we transform the problem for meeting the Markovian property, and model it as an infinite horizon semi-Markov decision process (SMDP). Achieving the delay optimality of the caching problem needs to solve the Bellman equation of the SMDP, but it leads to the curse of dimensionality. We decompose the global optimality equation into several per-content optimality equations, and propose a low-complexity delay-optimal content caching algorithm by stochastic learning for each content. Finally, the simulation results show that our proposed algorithm achieves significantly lower delay than conventional caching algorithms.

To adapt to the increasing storage demands and varying storage redundancy requirements, practical distributed storage systems need to supportstorage scalingby relocating currently stored data to different storage nodes. However, the scaling process inevitably transfers substantial data traffic over the network. Thus, minimizing the bandwidth cost of the scaling process is critical in distributed settings. In this paper, we show that optimal storage scaling is achievable in erasure-coded distributed storage based on network coding, by allowing storage nodes to send encoded data during scaling. We formally prove the information-theoretically minimum scaling bandwidth for both scale-out and scale-in cases. Based on our theoretical findings, we also build a distributed storage system prototype NCScale based on Hadoop Distributed File System, so as to realize network-coding-based scaling while preserving the necessary properties for practical deployment. Experiments on Amazon EC2 show that the scaling time can be reduced by up to 50% over the state-of-the-art.

More and more 5G and AI applications demand flexible and low-cost processing of their traffic through diverse virtualized network functions (VNFs) to meet their security and privacy requirements. As such, the Network Function Virtualization (NFV) market has been emerged as a major service market that allows network service providers to trade their network services among customers. Since each service market usually involves complex interplays among players with different roles, efficient mechanisms that guarantee stable and efficient operations of the NFV market are urgently needed. One fundamental problem in the NFV market is how to maximize the social welfare of all players so that all players have incentives to participate in the activities of the market. In this paper, we first formulate a novel social welfare maximization problem in an NFV market of a multi-tier edge cloud network, with the aim to maximize the total revenue collected from all players, and we implement VNF services on Virtual Machines (VMs) leased by service providers to fulfill customers with service requests, where the edge cloud network consists of both cloudlets in edge networks and remote data centers in the core network. We then design an efficient incentive-compatible mechanism for the problem, and analyze the existence of a Nash equilibrium of the mechanism. Also, we consider an online social welfare maximization problem with uncertain values of customers and without the knowledge of future request arrivals, for which we devise an online learning algorithm by adopting the Multi-Armed Bandits (MAB) method with a bounded regret. We finally evaluate the performance of the proposed mechanisms through simulations and a testbed. Results show that the proposed mechanisms deliver up to 27% higher social welfare than those of existing studies

Given a target binary function, the binary code search retrieves top-K similar functions in the repository, and similar functions represent that they are compiled from the same source codes. Searching binary code is particularly challenging due to large variations of compiler tool-chains and options and CPU architectures, as well as thousands of binary codes. Furthermore, there are some pivotal issues in current binary code search schemes, including inaccurate text-based or token-based analysis, slow graph matching, or complex deep learning processes. In this paper, we present an unsupervised tensor embedding scheme, Codee, to carry out code search efficiently and accurately at the binary function level. First, we use an NLP-based neural network to generate the semantic-aware token embedding. Second, we propose an efficient basic block embedding generation algorithm based on the network representation learning model. We learn both the semantic information of instructions and the control flow structural information to generate the basic block embedding. Then we use all basic block embeddings in a function to obtain a variable-length function feature vector. Third, we build a tensor to generate function embeddings based on the tensor singular value decomposition, which compresses the variable-length vectors into short fixed-length vectors to facilitate efficient search afterward. We further propose a dynamic tensor compression algorithm to incrementally update the function embedding database. Finally, we use the local sensitive hash method to find the top-$K$similar matching functions in the repository. Compared with state-of-the-art cross-optimization-level code search schemes, such as Asm2Vec and DeepBinDiff, our scheme achieves higher average search accuracy, shorter feature vectors, and faster feature generation performance using four datasets, OpenSSL, Coreutils, libgmp and libcurl. Compared with other cross-platform and cross-optimization-level code search schemes, such as Gemini, Safe, the average recall of our method also outperforms others.

This paper addresses the task of segmenting class-agnostic objects in semi-supervised setting. Although previous detection based methods achieve relatively good performance, these approaches extract the best proposal by a greedy strategy, which may lose the local patch details outside the chosen candidate. In this paper, we propose a novel spatiotemporal graph neural network (STG-Net) to reconstruct more accurate masks for video object segmentation, which captures the local contexts by utilizing all proposals. In the spatial graph, we treat object proposals of a frame as nodes and represent their correlations with an edge weight strategy for mask context aggregation. To capture temporal information from previous frames, we use a memory network to refine the mask of current frame by retrieving historic masks in a temporal graph. The joint use of both local patch details and temporal relationships allow us to better address the challenges such as object occlusions and missing. Without online learning and fine-tuning, our STG-Net achieves state-of-the-art performance on four large benchmarks, demonstrating the effectiveness of the proposed approach.

Although deep learning based methods have achieved great progress in unsupervised video object segmentation, difficult scenarios (e.g., visual similarity, occlusions, and appearance changing) are still no well-handled. To alleviate these issues, we propose a novel Focus on Foreground Network (F2Net), which delves into the intra-inter frame details for the foreground objects and thus effectively improve the segmentation performance. Specifically, our proposed network consists of three main parts: Siamese Encoder Module, Center Guiding Appearance Diffusion Module, and Dynamic Information Fusion Module. Firstly, we take a siamese encoder to extract the feature representations of paired frames (reference frame and current frame). Then, a Center Guiding Appearance Diffusion Module is designed to capture the inter-frame feature (dense correspondences between reference frame and current frame), intra-frame feature (dense correspondences in current frame), and original semantic feature of current frame. Different from the Anchor Diffusion Network, we establish a Center Prediction Branch to predict the center location of the foreground object in current frame and leverage the center point information as spatial guidance prior to enhance the inter-frame and intra-frame feature extraction, and thus the feature representation considerably focus on the foreground objects. Finally, we propose a Dynamic Information Fusion Module to automatically select relatively important features through three aforementioned different level features. Extensive experiments on DAVIS, Youtube-object, and FBMS datasets show that our proposed F2Net achieves the state-of-the-art performance with significant improvement.

This paper addresses the problem of temporal sentence grounding (TSG), which aims to identify the temporal boundary of a specific segment from an untrimmed video by a sentence query. Previous works either compare pre-defined candidate segments with the query and select the best one by ranking, or directly regress the boundary timestamps of the target segment. In this paper, we propose a novel localization framework that scores all pairs of start and end indices within the video simultaneously with a biaffine mechanism. In particular, we present a Context-aware Biaffine Localizing Network (CBLN) which incorporates both local and global contexts into features of each start/end position for biaffine-based localization. The local contexts from the adjacent frames help distinguish the visually similar appearance, and the global contexts from the entire video contribute to reasoning the temporal relation. Besides, we also develop a multi-modal self-attention module to provide fine-grained query-guided video representation for this biaffine strategy. Extensive experiments show that our CBLN significantly outperforms state-of-thearts on three public datasets (ActivityNet Captions, TACoS, and Charades-STA), demonstrating the effectiveness of the proposed localization framework. The code is available at https://github.com/liudaizong/CBLN.

We address the problem of temporal sentence localization in videos (TSLV).Traditional methods follow a top-down framework which localizes the target segment with predefined segment proposals.Although they have achieved decent performance, the proposals are handcrafted and redundant.Recently, bottom-up framework attracts increasing attention due to its superior efficiency.It directly predicts the probabilities for each frame as a boundary.However, the performance of bottom-up model is inferior to the top-down counterpart as it fails to exploit the segmentlevel interaction.In this paper, we propose an Adaptive Proposal Generation Network (APGN) to maintain the segment-level interaction while speeding up the efficiency.Specifically, we first perform a foregroundbackground classification upon the video and regress on the foreground frames to adaptively generate proposals.In this way, the handcrafted proposal design is discarded and the redundant proposals are decreased.Then, a proposal consolidation module is further developed to enhance the semantic of the generated proposals.Finally, we locate the target moments with these generated proposals following the top-down framework.Extensive experiments on three challenging benchmarks show that our proposed APGN significantly outperforms previous state-of-the-art methods.

A key solution to temporal sentence grounding (TSG) exists in how to learn effective alignment between vision and language features extracted from an untrimmed video and a sentence description.Existing methods mainly leverage vanilla soft attention to perform the alignment in a single-step process.However, such single-step attention is insufficient in practice, since complicated relations between inter-and intra-modality are usually obtained through multi-step reasoning.In this paper, we propose an Iterative Alignment Network (IA-Net) for TSG task, which iteratively interacts inter-and intra-modal features within multiple steps for more accurate grounding.Specifically, during the iterative reasoning process, we pad multi-modal features with learnable parameters to alleviate the nowhere-toattend problem of non-matched frame-word pairs, and enhance the basic co-attention mechanism in a parallel manner.To further calibrate the misaligned attention caused by each reasoning step, we also devise a calibration module following each attention module to refine the alignment knowledge.With such iterative alignment scheme, our IA-Net can robustly capture the fine-grained relations between vision and language domains step-bystep for progressively reasoning the temporal boundaries.Extensive experiments conducted on three challenging benchmarks demonstrate that our proposed model performs better than the state-of-the-arts.

Augmented Reality (AR) has various practical applications in healthcare, education, and entertainment. To provide a fully interactive and immersive experience, AR applications require extremely high responsiveness and ultra-low processing latency. Mobile edge computing (MEC) has shown great potential in meeting such stringent requirements and demands of AR applications by implementing AR requests in edge servers within the close proximity of these applications. In this paper, we investigate the problem of reward maximization for AR applications with uncertain demands in an MEC network, such that the reward of provisioning services for AR applications is maximized and the responsiveness of AR applications is enhanced, subject to both network resource capacity. We devise an exact solution for the problem if the problem size is small, otherwise we develop an efficient approximation algorithm with a provable approximation ratio for the problem. We also devise an online learning algorithm with a bounded regret for the dynamic reward maximization problem without the knowledge of the future arrivals of AR requests, by adopting the technique of Multi-Armed Bandits (MAB). We evaluate the performance of the proposed algorithms through simulations. Experimental results show that the proposed algorithms outperform existing studies by 17 % higher reward.

With the fast development of next-generation networking techniques, a Network Function Virtualization (NFV) market is emerging as a major market that allows network service providers to trade various network services among consumers. Therefore, efficient mechanisms that guarantee stable and efficient operations of the NFV market are urgently needed. One fundamental problem in the NFV market is how to maximize the social welfare of all players, so they have incentives to participate in activities of the market. In this paper, we first formulate the social welfare maximization problem, with an aim to maximize the total revenue of all players in the NFV market. For the social welfare maximization problem, we design an efficient incentive-compatible mechanism and analyze the existence of a Nash equilibrium of the mechanism. We also consider an online social welfare maximization problem without the knowledge of future request arrivals. We devise an online learning algorithm based on Multi-Armed Bandits (MAB) to allow both customers and network service providers to make decisions with uncertainty of customers' strategy. We evaluate the performance of the proposed mechanisms by both simulations and test-bed implementations, and the results show that the proposed mechanisms obtain at most 23% higher social welfare than existing studies.

Crowd counting has drawn much attention due to its importance in safety-critical surveillance systems. Especially, deep neural network (DNN) methods have significantly reduced estimation errors for crowd counting missions. Recent studies have demonstrated that DNNs are vulnerable to adversarial attacks, i.e., normal images with human-imperceptible perturbations could mislead DNNs to make false predictions. In this work, we propose a robust attack strategy called Adversarial Patch Attack with Momentum (APAM) to systematically evaluate the robustness of crowd counting models, where the attacker's goal is to create an adversarial perturbation that severely degrades their performances, thus leading to public safety accidents (e.g., stampede accidents). Especially, the proposed attack leverages the extreme-density background information of input images to generate robust adversarial patches via a series of transformations (e.g., interpolation, rotation, etc.). We observe that by perturbing less than 6% of image pixels, our attacks severely degrade the performance of crowd counting systems, both digitally and physically. To better enhance the adversarial robustness of crowd counting models, we propose the first regression model-based Randomized Ablation (RA), which is more sufficient than Adversarial Training (ADT) (Mean Absolute Error of RA is 5 lower than ADT on clean samples and 30 lower than ADT on adversarial examples). Extensive experiments on five crowd counting models demonstrate the effectiveness and generality of the proposed method.

Recent deep networks have convincingly demonstrated high capability in crowd counting, which is a critical task attracting widespread attention due to its various industrial applications. Despite such progress, trained data-dependent models usually can not generalize well to unseen scenarios because of the inherent domain shift. To facilitate this issue, this paper proposes a novel adversarial scoring network (ASNet) to gradually bridge the gap across domains from coarse to fine granularity. In specific, at the coarse-grained stage, we design a dual-discriminator strategy to adapt source domain to be close to the targets from the perspectives of both global and local feature space via adversarial learning. The distributions between two domains can thus be aligned roughly. At the fine-grained stage, we explore the transferability of source characteristics by scoring how similar the source samples are to target ones from multiple levels based on generative probability derived from coarse stage. Guided by these hierarchical scores, the transferable source features are properly selected to enhance the knowledge transfer during the adaptation process. With the coarse-to-fine design, the generalization bottleneck induced from the domain discrepancy can be effectively alleviated. Three sets of migration experiments show that the proposed methods achieve state-of-the-art counting performance compared with major unsupervised methods.

By caching popular contents at edge nodes, the content can be delivered via device-to-device (D2D) communications to reduce the delay significantly. In this paper, we study cooperative content delivery from multiple edge nodes with duplicate caching, where the content caching and the cooperative transmission algorithms are designed jointly to minimize the average delay. Specifically, we decompose the delay optimization problem as short time-scale transmission subproblems for each time slot and a long time-scale caching master problem by primal decomposition. The transmission subproblem is modelled as a finite horizon Markov decision problem (MDP), and we propose a link selection algorithm based on the value function in the MDP model. For the caching master problem, we propose a caching algorithm by the simultaneous perturbation stochastic approximation (SPSA) method. Simulation results show that our proposed transmission and caching algorithms achieve lower delay than the conventional schemes.

Internet of Underwater Things (IoUT) has a wide range of application prospects in civil and military fields. As a key enabling technology of IoUT, underwater acoustic sensor networks (UASNs) feature a variety of unique characteristics, including high latency, high mobility, and low bandwidth. All these problems pose challenges in the design of efficient and effective routing protocols. To address such challenges, we propose an adaptive-location-based routing protocol named ALRP for 3-D UASNs. First, the ALRP defines a forwarding area to confine the scope of the candidate forwarders. Second, it adaptively calculates the forwarding probability to reduce redundant forwarding. Third, it adaptively obtains the forwarding delay to determine the forwarding order. By using these ways, ALRP can find the beneficial relays to forward packets, and increase the forwarding efficiency. We implement the ALRP and evaluate its performance in the Aqua-sim. Simulation results show that the ALRP significantly improves the network performance compared to several existing protocols.

Internet-of-Things-enabled E-health system, which could monitor and collect the personal health information (PHI), has gradually transformed the clinical treatment to a more personalized way with in-home monitoring smart devices. Then, with the collected PHI, clinical decision support systems (CDSSs), which are based on data mining techniques and historical electronic medical records (EMRs) to help clinicians make proper treatment decisions, have attracted considerable attention. To address issues, such as network congestion and low rate of responsiveness for traditional methods when implementing CDSSs, we integrate the technologies mobile-edge computing (MEC) and software-defined networking for exploiting the computation resources and storage capacities among edge nodes (ENs) (i.e., MEC servers) in our model. Based on this integrated system, each edge node will deploy a double deep Q-network (DDQN) to obtain a stable and sequential clinical treatment policy. It is enabled by a novel fully decentralized federated framework (FDFF) for aggregating models of DDQN and extracting the knowledge from EMRs across all ENs. Furthermore, we discuss the convergence of FDFF in resource-constrained environments. However, since most EMRs are faced with stringent privacy concerns, we adopt two additively homomorphic encryption schemes to prevent leakage of EMRs' privacy during the training process of FDFF. Finally, we measure the time cost of our additively homomorphic encryption schemes and validate DDQN with experiments on large data sets based on FDFF, which shows promising performance on clinician treatment.

Abstract Deep learning techniques have achieved remarkable performance in wide-ranging tasks. However, when trained on privacy-sensitive datasets, the model parameters may expose private information in training data. Prior attempts for differentially private training, although offering rigorous privacy guarantees, lead to much lower model performance than the non-private ones. Besides, different runs of the same training algorithm produce models with large performance variance. To address these issues, we propose DPlis– Differentially Private Learning wIth Smoothing. The core idea of DPlis is to construct a smooth loss function that favors noise-resilient models lying in large flat regions of the loss landscape. We provide theoretical justification for the utility improvements of DPlis. Extensive experiments also demonstrate that DPlis can effectively boost model quality and training stability under a given privacy budget.

Mobile edge computing becomes a promising technology to mitigate the latency of various cloud services. In addition, network function virtualization (NFV) has been shown a great potential in reducing the operational cost of cloud services while enhancing the flexibility of virtual network function deployments, by implementing dedicated hardware network functions as pieces of software in generic servers. Recently, the GPU acceleration has been investigated to speed up flow processing in virtual network functions (VNFs), by leveraging the parallelism of GPUs. VNFs that need accelerations prefer to stay at cloudlets (locations) equipped with GPUs. However, little attention has been paid for the VNF placement that takes into account GPU-affinity in cloudlets of mobile edge clouds. In this paper, we consider the affinity-aware throughput maximization problem in a mobile edge cloud via leveraging the parallelism on GPUs for user requests with VNF requirements. We consider two types of affinities in the VNF placement: Thesoft-affinitythat allows VNFs to be executed by either CPUs or GPUs in cloudlets; and thehard-affinitythat only allows VNFs to be placed to the GPUs of a specified set of cloudlets. We formulate two corresponding VNF placement problems in a mobile edge cloud. Specifically, we first propose an exact solution to the soft-affinity throughput maximization problem by formulating an Integer Linear Program (ILP). We then propose an efficient algorithm for the problem, by proposing a randomized algorithm with a provable approximation ratio for the hard-affinity-aware throughput maximization problem and extending the proposed approximation algorithm to the soft-affinity throughput maximization problem. Furthermore, assuming that user requests arrive into the mobile edge cloud one by one without the knowledge of future arrivals, we devise an online algorithm with a good competitive ratio for this dynamic hard-affinity-aware throughput maximization problem. Finally, we evaluate the performance of the proposed algorithms, through simulations and implementations in a real test-bed. Experimental results show that the performance of the proposed algorithms outperform their existing counterparts and achieve higher throughput.

In this article, we propose a memristor-based ShuffleNetV2 for image classification. Because of the low power consumption and high integration, this circuit is suitable for edge computing. The memristor-based ShuffleNetV2 is divided into four kinds of units, and each unit is composed by a series of basic memristive neural circuits, such as memristive convolutional neural networks (MCNNs), memristive batch normalization (MBN) layers, memristive fully connection (MFC) layers, rectified linear unit (ReLU) layers, max-pooling layers, etc. To solve the imbalance of required memristor crossbars in MCNN and the huge power consumption of the input voltage inverting, the inverters are shifted behind the memristor crossbar.This circuit uses (M+1)×2 memristor crossbars instead of the original (2M+1)×1 memristor crossbars. The world lines of memristor crossbars required can be reduced from (2M+1) to (M+1). The number of inverters in MCNN for a kernel can be decreased to one, and the power consumption in MCNN is greatly decreased. The MBN is designed by combining operational amplifiers with memristors. An output limiter circuit is added behind the ReLU layer implemented by diode. The output voltage of the ReLU layer is limited to the threshold voltage Vpof memristor. The memristive ShuffleNetV2 is established for image classification in experiments. The effectiveness of the circuit is verified on the FER2013 dataset along with the analysis of the memristor resources, calculation periods, and circuit power consumption of the network. The experimental results show that this that the inference time of this memristive ShuffleNetV2 could be as less as 137 ns with a max power consumption 2.8 μW of a memristive neurons.

The ocean has been investigated for centuries across the world, and planning the travel path for vessels in the ocean has become a hot topic in recent decades as the increasing development of worldwide business trading. Planning such suitable paths is often based on big data processing in cybernetics, while not many investigations have been done. We attempt to find the optimal path for vessels in the ocean by proposing an online learning dispatch approach on studying the mission-executing-feedback (MEF) model. The proposed approach explores the ocean subdomain (OS) to achieve the largest average traveling feedback for different vessels. It balances the ocean path by a deep and wide search, and considers adaptation for these vessels. Further, we propose a contextual multiarmed bandit-based algorithm, which provides accurate exploration results with sublinear regret and significantly improves the learning speed. The experimental results show that the proposed MEF approach possesses 90% accuracy gain over random exploration and achieves about 25% accuracy improvement over other contextual bandit models on supporting big data online learning pre-eminently.

One existing challenge associated with large scale skyline queries on cloud services, particularly when dealing with private information such as biomedical data, is supporting multi-party queries with curious-but-honest parties on encrypted data. In addition, existing solutions designed for performing secure skyline queries incur significant communication and computation costs due to ciphertext calculation. Thus, in this paper, we demonstrate the potential of supporting privacy-preserving multi-party skyline queries on encrypted data using additive homomorphic and proxy re-encryption cryptosystems. However, the secure computation based on these cryptosystems will further slow down query efficiency. To improve the efficiency of comparison on encrypted data, we redesign two lightweight secure comparison protocols. Meanwhile, we present an efficient method named “blind-reading” to securely obtain the skyline point. We also propose a novel method, Privacy Matrix, designed to reduce the scale of the dataset so that the computational cost is significantly decreased without privacy leakage. Then, we construct our secure skyline query protocol by integrating lightweight secure comparison protocols, “blind-reading” and Privacy Matrix techniques. Finally, we evaluate the security of our protocol, where we show it is secure without leaking information. The performance evaluation also shows that our proposed approach significantly improves the efficiency (at least ×4.5 faster) compared to the-state-of-art and has the scalability of query processing under large datasets.

Deep learning-based methods have achieved remarkable success in image restoration and enhancement, but are they still competitive when there is a lack of paired training data? As one such example, this paper explores the low-light image enhancement problem, where in practice it is extremely challenging to simultaneously take a low-light and a normal-light photo of the same visual scene. We propose a highly effective unsupervised generative adversarial network, dubbed EnlightenGAN, that can be trained without low/normal-light image pairs, yet proves to generalize very well on various real-world test images. Instead of supervising the learning using ground truth data, we propose to regularize the unpaired training using the information extracted from the input itself, and benchmark a series of innovations for the low-light image enhancement problem, including a global-local discriminator structure, a self-regularized perceptual loss fusion, and the attention mechanism. Through extensive experiments, our proposed approach outperforms recent methods under a variety of metrics in terms of visual quality and subjective user study. Thanks to the great flexibility brought by unpaired training, EnlightenGAN is demonstrated to be easily adaptable to enhancing real-world images from various domains. Our codes and pre-trained models are available at: https://github.com/VITA-Group/EnlightenGAN.

Autonomous on-demand services, such as GOGOX (formerly GoGoVan) in Hong Kong, provide a platform for users to request services and for suppliers to meet such demands. In such a platform, the suppliers have autonomy to accept or reject the demands to be dispatched to him/her, so it is challenging to make an online matching between demands and suppliers. Existing methods use round-based approaches to dispatch demands. In these works, the dispatching decision is based on the predicted response patterns of suppliers to demands in the current round, but they all fail to consider the impact of future demands and suppliers on the current dispatching decision. This could lead to taking a suboptimal dispatching decision from the future perspective. To solve this problem, we propose a novel demand dispatching model using deep reinforcement learning. In this model, we make each demand as an agent. The action of each agent, i.e., the dispatching decision of each demand, is determined by a centralized algorithm in a coordinated way. The model works in the following two steps. (1) It learns the demand’s expected value in each spatiotemporal state using historical transition data. (2) Based on the learned values, it conducts a Many-To-Many dispatching using a combinatorial optimization algorithm by considering both immediate rewards and expected values of demands in the next round. In order to get a higher total reward, the demands with a high expected value (short response time) in the future may be delayed to the next round. On the contrary, the demands with a low expected value (long response time) in the future would be dispatched immediately. Through extensive experiments using real-world datasets, we show that the proposed model outperforms the existing models in terms of Cancellation Rate and Average Response Time.

Nowadays, the booming demand of big data analytics and the constraints of computational ability and network bandwidth have made it difficult for a stand-alone agent/service provider to provide suitable information for every user from the large volume online data within the limited time. To handle this challenge, a recommender system (RS) can call in a group of agents to collaborate to learn users' preference and taste, which is known as a distributed recommender system (DRS). DRSs can improve the accuracy of a traditional RS by requesting agents to share information with each other. However, it is challenging for DRSs to make personalized recommendations for each user due to the large amount of candidates. In addition, information sharing among agents raises a privacy concern. Thus, we propose a privacy-preserving DRS in this paper, and then model each service provider as a distributed online learner with context-awareness. Service providers collaborate to make personalized recommendations by learning users' preferences according to the user context and users' history behaviors. We adopt the federated learning framework to help train a high quality privacy- preserving centralized model over a large number of distributed agents which is probably unreliable with relatively slow network connections. To handle big data scenario, we build an item-cluster tree to deal with online and increasing datasets from top to the bottom. We further consider the structure of social network and present an efficient algorithm to avoid more performance loss adaptively. Theoretical proofs show that our proposed algorithm can achieve sublinear regret and differential privacy protection simultaneously for service providers and users. Numerical results confirm that our novel framework can handle increasing big datasets and strike a trade-off between privacy-preserving level and the prediction accuracy.

Mobile Edge Computing (MEC) pushes computing functionalities away from the centralized cloud to the proximity of data sources, thereby reducing service provision latency and saving backhaul network bandwidth. Although computation offloading for MEC systems has been extensively studied in the literature, service placement is an equally, if not more, important design topic of MEC, yet receives much less attention. Service placement refers to configuring the service platform and storing the related libraries/databases at the edge server, e.g., MEC-enabled Base Station (BS), which enables corresponding computation tasks to be executed. Due to the limited computing resource, the edge server can host only a small number of services and hence which services to host has to be judiciously decided to maximize the system performance. In this paper, we investigate collaborative service placement in MEC-enabled dense small cell networks. An efficient decentralized algorithm, called CSP (Collaborative Service Placement), is proposed where a network of small cell BSs optimize service placement decisions collaboratively to address a number of challenges in MEC systems, including service heterogeneity, spatial demand coupling, and decentralized coordination. CSP is developed based on parallel Gibbs sampling by exploiting the graph coloring on the small cell network. The algorithm significantly improves the time efficiency compared to conventional Gibbs sampling, yet guarantees provable convergence and optimality. CSP is further extended to work with selfish BSs, where BSs are allowed to choose “to cooperate” or “not to cooperate.” We employ coalitional game to investigate the strategic behaviors of selfish BSs and design a coalition formation scheme to form stable BS coalitions using merge-and-split rules. Simulations results show that CSP can effectively reduce edge system operational cost for both cooperative and selfish BSs.

In mobile crowdsensing, it is a challenge to assign tasks to appropriate smartphones. Existing task allocation mechanisms mainly aim at optimizing the global system performance, while ignoring the personal preferences of individual crowdsensing tasks and smartphone users. Nevertheless, in an open crowdsensing system, a task assignment is prone to be unstable if smartphone users or tasks have incentives to deviate from the global assignment, and seek for alternative choices to improve their own utilities. Besides that, during task competition, the rational smartphone users might choose to adjust their payments after the first few failures, which however, brings new challenges in achieving the stability. To address these issues, this paper constructs a distributed many-to-many matching model to capture the interaction between crowdsensing tasks and smartphone users, taking into account the budget constraints of tasks. Then, we design a stable matching algorithm to allocate the tasks to the users, and determine their payments. We prove that the proposed algorithm achieves several desirable properties including individual rationality, stability, and convergency. It is also proved that the proposed scheme achieves at least half of the optimal system efficiency when each smartphone provides homogeneous service quality. Finally, simulation results confirm the effectiveness of the proposed scheme.

Recent techniques built on generative adversarial networks (GANs), such as cycle-consistent GANs, are able to learn mappings among different domains built from unpaired data sets, through min-max optimization games between generators and discriminators. However, it remains challenging to stabilize the training process and thus cyclic models fall into mode collapse accompanied by the success of discriminator. To address this problem, we propose an novel Bayesian cyclic model and an integrated cyclic framework for interdomain mappings. The proposed method motivated by Bayesian GAN explores the full posteriors of cyclic model via sampling latent variables and optimizes the model with maximum a posteriori (MAP) estimation. Hence, we name it Bayesian CycleGAN. In addition, original CycleGAN cannot generate diversified results. But it is feasible for Bayesian framework to diversify generated images by replacing restricted latent variables in inference process. We evaluate the proposed Bayesian CycleGAN on multiple benchmark data sets, including Cityscapes, Maps, and Monet2photo. The proposed method improve the per-pixel accuracy by 15% for the Cityscapes semantic segmentation task within origin framework and improve 20% within the proposed integrated framework, showing better resilience to imbalance confrontation. The diversified results of Monet2Photo style transfer also demonstrate its superiority over original cyclic model. We provide codes for all of our experiments in https://github.com/ranery/Bayesian-CycleGAN.

The consistency of a response to a given post at the semantic level and emotional level is essential for a dialogue system to deliver humanlike interactions. However, this challenge is not well addressed in the literature, since most of the approaches neglect the emotional information conveyed by a post while generating responses. This article addresses this problem and proposes a unified end-to-end neural architecture, which is capable of simultaneously encoding the semantics and the emotions in a post and leveraging target information to generate more intelligent responses with appropriately expressed emotions. Extensive experiments on real-world data demonstrate that the proposed method outperforms the state-of-the-art methods in terms of both content coherence and emotion appropriateness.

With increasing focus on Artificial Intelligence (AI) applications, Deep Neural Networks (DNNs) have been successfully used in a number of application areas. As the number of layers and neurons in DNNs increases rapidly, significant computational resources are needed to execute a learned DNN model. This ever-increasing resource demand of DNNs is currently met by large-scale data centers with state-of-the-art GPUs. However, increasing availability of mobile edge computing and 5G technologies provide new possibilities for DNN-driven AI applications, especially where these application make use of data sets that are distributed in different locations. One fundamental process of a DNN-driven application in mobile edge clouds is the adoption of “inferencing” - the process of executing a pre-trained DNN based on newly generated image and video data from mobile devices. We investigate offloading DNN inference requests in a 5G-enabled mobile edge cloud (MEC), with the aim to admit as many inference requests as possible. We propose exact and approximate solutions to the problem of inference offloading in MECs. We also consider dynamic task offloading for inference requests, and devise an online algorithm that can be adapted in real time. The proposed algorithms are evaluated through large-scale simulations and using a real world test-bed implementation. The experimental results demonstrate that the empirical performance of the proposed algorithms outperform their theoretical counterparts and other similar heuristics reported in literature.

This article presents a simple and intuitive solution for multilabel image classification, which achieves the competitive performance on the popular COCO and PASCAL VOC benchmarks. The main idea is to capture how humans perform this task: we recognize both labels (i.e., objects and attributes) and the correlation of labels at the same time. Here, label recognition is performed by a standard ConvNet pipeline, whereas label correlation modeling is done by projecting both labels and image features extracted by the ConvNet to a common latent vector space. Specifically, we carefully design the loss function to ensure that: 1) labels and features that co-appear frequently are close to each other in the latent space and 2) conversely, labels/features that do not appear together are far apart. This information is then combined with the original ConvNet outputs to form the final prediction. The whole model is trained end-to-end, with no additional supervised information other than the image-level supervised information. Experiments show that the proposed method consistently outperforms previous approaches on COCO and PASCAL VOC in terms of mAP, macro/micro precision, recall, and$F$-measure. Further, our model is highly efficient at test time, with only a small number of additional weights compared to the base model for direct label recognition.

Temporal sentence localization in videos aims to ground the best matched segment in an untrimmed video according to a given sentence query.Previous works in this field mainly rely on single-step attentional frameworks to align the temporal boundaries by a soft selection.Although they focus on the visual content relevant to the query, these attention strategies are insufficient to model complex video contents and restrict the higher-level reasoning demand for temporal relation.In this paper, we propose a novel deep rectification-modulation network (RMN), transforming this task into a multi-step reasoning process by repeating rectification and modulation.In each rectification-modulation layer, unlike existing methods directly conducting the cross-modal interaction, we first devise a rectification module to correct implicit attention misalignment which focuses on wrong position during the interaction process.Then, a modulation module is developed to model the frame-to-frame relation with the help of specific sentence information for better correlating and composing the video contents over time.With multiple such layers cascaded in depth, our RMN progressively refines video and query interactions, thus enabling a further precise localization.Experimental evaluations on three public datasets show that the proposed method achieves state-of-the-art performance.

The task of crowd counting is extremely challenging due to complicated difficulties, especially the huge variation in vision scale.Previous works tend to adopt a naive concatenation of multiscale information to tackle it, while the scale shifts between the feature maps are ignored.In this paper, we propose a novel Hierarchical Scale Recalibration Network (HSRNet), which addresses the above issues by modeling rich contextual dependencies and recalibrating multiple scale-associated information.Specifically, a Scale Focus Module (SFM) first integrates global context into local features by modeling the semantic inter-dependencies along channel and spatial dimensions sequentially.In order to reallocate channel-wise feature responses, a Scale Recalibration Module (SRM) adopts a step-bystep fusion to generate final density maps.Furthermore, we propose a novel Scale Consistency loss to constrain that the scale-associated outputs are coherent with groundtruth of different scales.With the proposed modules, our approach can ignore various noises selectively and focus on appropriate crowd scales automatically.Extensive experiments on crowd counting datasets (ShanghaiTech, MALL, WorldEXPO'10, and UCSD) show that our HSRNet can deliver superior results over all state-of-the-art approaches.More remarkably, we extend experiments on an extra vehicle dataset , whose results indicate that the proposed model is generalized to other applications.

Leveraging the content caching technology, popular contents can be in close proximity to the users, which reduces the delay effectively. In this paper, we dedicate to minimizing the long-term average delay for the cache-enabled wireless networks, in which the fetching delay is considered when the uncached contents are retrieved from remote servers. We construct a decision-theoretic framework for this delay-optimal content-caching problem. The most challenging obstacle is that the consideration of imperfect content fetching breaks the Markovian property in the decision-theoretic framework, which makes the problem difficult to solve. To overcome this obstacle, we propose an equivalent transformation by analyzing the queue dynamics during the content fetching delay, so that the transformed problem can be modeled as an infinite horizon semi-Markov decision process (SMDP). To handle the curse of dimensionality for solving the SMDP, we decompose the global optimality equation into several per-content equations, and propose a low-complexity delay-optimal content caching algorithm. Simulation results show that the proposed algorithm achieves significantly lower delay than conventional caching algorithms.

Mobile edge computing (MEC) is emerging as an enabling technology of low-latency network services, such as Augmented Reality (AR) and Virtual Reality (VR), by deploying cloudlets in locations close to users. In MEC networks, telcooperators can place their services to cloudlets, such that the service accessing delay of users is minimized. In this paper, we investigate a fundamental problem of caching services that are originally deployed in remote clouds to cloudlets in an MEC network within the proximity of users. Specifically, we focus on the service caching problem in a two-tiered MEC network with both remote clouds and cloudlets that are close to users, in which multiple network service providers competing computing and bandwidth resources. This setting is significantly different from existing studies that focused on offloading user tasks from mobile devices to cloudlets in MEC networks that typically do not consider a service market with multiple network service providers. For the service caching problem in a two-tiered MEC network, we propose a novel approximation-restricted framework that guarantees the stableness of the service market. Under the proposed framework, an approximation algorithm with an approximation ratio for the problem with non-selfish players and an efficient, stable Stackelberg congestion game with selfish players have been proposed. We also analyze the Price of Anarchy (PoA) of the proposed Stackelberg congestion game to measure the efficiency of the proposed game degrades due to selfish behavior of network service providers. We finally evaluate the performance of our mechanism on both simulated environments and a real test-bed. Results show that the performance of our proposed mechanism is promising.

In this paper, we study the privacy of online health data. We present a novel online health data De-Anonymization (DA) framework, named De-Health. Leveraging two real world online health datasets WebMD and HealthBoards, we validate the DA efficacy of De-Health. We also present a linkage attack framework which can link online health/medical information to real world people. Through a proof-of-concept attack, we link 347 out of 2805 WebMD users to real world people, and find the full names, medical/health information, birthdates, phone numbers, and other sensitive information for most of the re-identified users. This clearly illustrates the fragility of the privacy of those who use online health forums.

Facial expression recognition (FER), aiming to classify the expression present in the facial image or video, has attracted a lot of research interests in the field of artificial intelligence and multimedia. In terms of video based FER task, it is sensible to capture the dynamic expression variation among the frames to recognize facial expression. However, existing methods directly utilize CNN-RNN or 3D CNN to extract the spatial-temporal features from different facial units, instead of concentrating on a certain region during expression variation capturing, which leads to limited performance in FER. In our paper, we introduce a Graph Convolutional Network (GCN) layer into a common CNN-RNN based model for video-based FER. First, the GCN layer is utilized to learn more significant facial expression features which concentrate on certain regions after sharing information between extracted CNN features of nodes. Then, a LSTM layer is applied to learn long-term dependencies among the GCN learned features to model the variation. In addition, a weight assignment mechanism is also designed to weight the output of different nodes for final classification by characterizing the expression intensities in each frame. To the best of our knowledge, it is the first time to use GCN in FER task. We evaluate our method on three widely-used datasets, CK+, Oulu-CASIA and MMI, and also one challenging wild dataset AFEW8.0, and the experimental results demonstrate that our method has superior performance to existing methods.

An essential yet challenging issue in crowd counting is the diverse background variations under complicated real-life environments, which makes attention based methods favorable in recent years. However, most existing methods only rely on first-order attention schemes (e.g. 2D position-wise attention), while ignoring the higher-order information within the congested scenes completely. In this paper, we propose a hybrid attention-aware network (HANet) with a high-order attention module (HAM) and an adaptive compensation loss (ACLoss) to tackle this problem. On the one hand, the HAM applies 3D attention to capture the subtle discriminative features around each people in the crowd. On the other hand, with the distributed supervision, the ACLoss exploits the prior knowledge from higher-level stages to guide the density map prediction at a lower level. The proposed HANet is then established with HAM and ACLoss working as different roles and promoting each other. Extensive experimental results show the superiority of our HANet against the state-of-the-arts on three challenging benchmarks.

Audio adversarial examples, imperceptible to humans, have been constructed to attack automatic speech recognition (ASR) systems. However, the adversarial examples generated by existing approaches usually incorporate noticeable noises, especially during the periods of silences and pauses. Moreover, the added noises often break temporal dependency property of the original audio, which can be easily detected by state-of-the-art defense mechanisms. In this paper, we propose a new Iterative Proportional Clipping (IPC) algorithm that preserves temporal dependency in audios for generating more robust adversarial examples. We are motivated by an observation that the temporal dependency in audios imposes a significant effect on human perception. Following our observation, we leverage a proportional clipping strategy to reduce noise during the low-intensity periods. Experimental results and user study both suggest that the generated adversarial examples can significantly reduce human-perceptible noises and resist the defenses based on the temporal structure.

To cope with the exploding mobile traffic in the fifth generation cellular network, the dense deployment of small cells and cognitive radios are two key technologies that significantly increase the network capacity and improve the spectrum utilization efficiency. Despite the desirable features, small cell cognitive radio networks (SCRNs) also face a higher risk of unauthorized spectrum access, which should not be overlooked. In this paper, we consider a passive monitoring system for SCRNs, which deploys sniffers for wireless traffic capture and network forensics, and study the optimal sniffer channel assignment (SCA) problem to maximize the monitoring performance. Unlike most existing SCA approaches that concentrate on user activity, we highlight the inherent error in wireless data capture (i.e. imperfect monitoring) due to the unreliable nature of wireless propagation, and propose an online-learning based algorithm called OSA (Online Sniffer-channel Assignment). OSA is a type of contextual combinatorial multi-armed bandit learning algorithm, which addresses key challenges in SCRN monitoring including the time- varying spectrum resource, imperfect monitoring, and uncertain network conditions. We theoretically prove that OSA has a sublinear learning regret bound and illustrate via simulations that OSA significantly outperforms benchmark solutions.

With the development of 5G technology, mobile edge computing is emerging as an enabling technique to promote Quality of Service (QoS) of network services. In particular, the response latency of network services can be significantly reduced by deploying cloudlets at 5G base stations in mobile edge clouds. Network service providers that usually deploy their services in remote clouds now shift their services from the remote clouds to the network edge in the proximity of users. However, the permanent placement of their services into edge clouds may not be economic, since computing and bandwidth resources in edge clouds are limited and relatively expensive. A smart way is to cache the services that are frequently requested by mobile users in edge clouds. In this paper, we study the problem of service caching in mobile edge network under a mobile service market with multiple network service providers completing for both computation and bandwidth resources of the edge cloud. We propose an Integer Linear Program (ILP) and a randomized rounding algorithm, for the problem without resource sharing among the network service providers. We also devise a distributed and stable game-theoretical mechanism for the problem with resource sharing among the network service providers, with the objective to minimize the social cost of all network service providers, by introducing a novel cost sharing model and a coalition formation game. We analyze the performance of the mechanism by showing a good guaranteed gap between the solution obtained and the optimal one, i.e., Strong Price of Anarchy (SPoA). We finally evaluate the performance of our algorithms by extensive simulations, and the obtained results show that the social cost of all players can be reduced significantly via allowing cooperation among network service providers in service caching.

Given high-speed mobile Internet access today, audiences are expecting much higher video quality than before. Video service providers have deployed dynamic video bitrate adaptation services to fulfill such user demands. However, legacy video bitrate adaptation techniques are highly dependent on the estimation of dynamic bandwidth, and fail to integrate the video quality enhancement techniques, or consider the heterogeneous computing capabilities of client devices, leading to low quality of experience (QoE) for users. In this paper, we present a super-resolution based adaptive video streaming (SRAVS) framework, which applies a Reinforcement Learning (RL) model for integrating the video super-resolution (VSR) technique with the video streaming strategy. The VSR technique allows clients to download low bitrate video segments, reconstruct and enhance them to high-quality video segments while making the system less dependent on estimating dynamic bandwidth. The RL model investigates both the playback statistics and the distinguishing features related to the client-side computing capabilities. Trace-driven emulations over real-world videos and bandwidth traces verify that SRAVS can significantly improve the QoE for users compared to the state-of-the-art video streaming strategies with or without involving VSR techniques.

Query-based moment localization is a new task that localizes the best matched segment in an untrimmed video according to a given sentence query. In this localization task, one should pay more attention to thoroughly mine visual and linguistic information. To this end, we propose a novel Cross- and Self-Modal Graph Attention Network (CSMGAN) that recasts this task as a process of iterative messages passing over a joint graph. Specifically, the joint graph consists of Cross-Modal interaction Graph (CMG) and Self-Modal relation Graph (SMG), where frames and words are represented as nodes, and the relations between cross- and self-modal node pairs are described by an attention mechanism. Through parametric message passing, CMG highlights relevant instances across video and sentence, and then SMG models the pairwise relation inside each modality for frame (word) correlating. With multiple layers of such a joint graph, our CSMGAN is able to effectively capture high-order interactions between two modalities, thus enabling a further precise localization. Besides, to better comprehend the contextual details in the query, we develop a hierarchical sentence encoder to enhance the query understanding. Extensive experiments on four public datasets demonstrate the effectiveness of our proposed model, and GCSMAN significantly outperforms the state-of-the-arts.

Temporal language localization in videos aims to ground one video segment in an untrimmed video based on a given sentence query. To tackle this task, designing an effective model to extract ground-ing information from both visual and textual modalities is crucial. However, most previous attempts in this field only focus on unidirectional interactions from video to query, which emphasizes which words to listen and attends to sentence information via vanilla soft attention, but clues from query-by-video interactions implying where to look are not taken into consideration. In this paper, we propose a Fine-grained Iterative Attention Network (FIAN) that consists of an iterative attention module for bilateral query-video in-formation extraction. Specifically, in the iterative attention module, each word in the query is first enhanced by attending to each frame in the video through fine-grained attention, then video iteratively attends to the integrated query. Finally, both video and query information is utilized to provide robust cross-modal representation for further moment localization. In addition, to better predict the target segment, we propose a content-oriented localization strategy instead of applying recent anchor-based localization. We evaluate the proposed method on three challenging public benchmarks: ActivityNet Captions, TACoS, and Charades-STA. FIAN significantly outperforms the state-of-the-art approaches.

With the development of deep learning technologies, attribute recognition and person re-identification (re-ID) have attracted extensive attention and achieved continuous improvement via executing computing-intensive deep neural networks in cloud datacenters. However, the datacenter deployment cannot meet the real-time requirement of attribute recognition and person re-ID, due to the prohibitive delay of backhaul networks and large data transmissions from cameras to datacenters. A feasible solution thus is to employ mobile edge clouds (MEC) within the proximity of cameras and enable distributed inference.

Machine learning (ML) has been penetrating into our daily life by facilitating many daily applications, e.g., self-driving, cloud gaming, product fault detection and drones. Meanwhile, there is an emerging trend that adopts ML methods into network optimization problems, such as flow classification, traffic engineering, routing, and etc. Conventional ML methods need careful training for a specific application of a given network structure, and the trained model normally cannot be applied to other applications and network structures. In this paper, we aim to design adaptive ML methods for network optimization problems, with the trained models having the ability of being deployed to any similar problems. In particular, we consider the virtualized network function (VNF) provisioning problem as our target optimization problem. We first propose a deep Q-learning-based optimization framework for VNF provisioning in a mobile edge network with network capacity constraints, by devising an adaptive graph feature embedding method. We then propose a series of deep Q-learning based learning algorithms for the problems of service chaining and the throughput maximization, based on the proposed learning-based optimization framework. We also propose a novel design of master-slave dual neural network that enables the decisions on both cloudlet selections and routing path finding. To stabilize and accelerate the convergence of the proposed methods, we devise a novel environment generation and termination strategy and a new structure for the replay buffer. We also evaluate the performance of the proposed framework and algorithms by extensive simulations. Results show that the proposed algorithms outperform existing methods by around 12%, and the trained model in a network can be directly adapted to other network structures and settings.

Industry 4.0 aims to create a modern industrial system by introducing technologies, such as cloud computing, intelligent robotics, and wireless sensor networks. In this article, we consider the multichannel access and task offloading problem in mobile-edge computing (MEC)-enabled industry 4.0 and describe this problem in multiagent environment. To solve this problem, we propose a novel multiagent deep reinforcement learning (MADRL) scheme. The solution enables edge devices (EDs) to cooperate with each other, which can significantly reduce the computation delay and improve the channel access success rate. Extensive simulation results with different system parameters reveal that the proposed scheme could reduce computation delay by 33.38% and increase the channel access success rate by 14.88% and channel utilization by 3.24% compared to the traditional single-agent reinforcement learning method.

The extensive applications in the Internet of Things (IoT) have inspired a growing network scale. However, due to the resource-limited IoT devices and the numerous cyber attacks against applications, maintaining the robustness and communication capabilities for the applications is increasingly challenging. In this article, we consider IoT network topologies that provide robust communication for heterogeneous networks and study the networking stability of IoT devices and the intelligent evolution computing in network architectures. We explicate the network robustness problem both for the network architecture and the resistance to cyber attacks. For the network architecture, we optimize the robustness of IoT network topology with a scale-free network model which has good performance in random attacks. In the case with the resistance to cyber attacks, a deep deterministic learning policy (DDLP) algorithm is proposed to improve the stability for large-scale IoT applications. Simulations show that the proposed algorithms greatly advance the robustness of IoT network topology compared to other algorithms, with a less computational cost.

The increasing Internet-of-Things (IoT) devices have produced large volumes of data. A deep learning technique is widely used to analyze the potential value of these data due to its unprecedented performance in both the academic and industrial communities. However, the data generated from the IoT devices are distributed among different users. Directly combining these data to a central server will cause privacy leakage, especially for personal sensitive data. Rather than centralized training by getting access to all these raw data, an alternative is to collaboratively learn a model in a distributed manner. However, there exist two main challenges in a distributed learning setting. The first one is how to preserve the privacy of users. The second one is to reduce the communication burden (e.g., mobile users have limited bandwidth) due to high-frequent data exchange. To address these two challenges, we design a communication efficient and privacy-preserving framework to enable different participants to distributively learn a model with a privacy protection guarantee. In particular, we develop a differentially private approximate mechanism for the distributed deep learning. In addition, we design a new gradient sparsification method to, at the first time, reduce both upload and download communication costs. The performance of the proposed framework is tested under different neural network structures for different data sets including, image classification and mobile sensor data. The experimental results demonstrate that we can reduce the communication up to only 2% compared to the full gradients exchange and achieve up to 16% accuracy increase compared to the previous works.

Telecommunication networks are undergoing a disruptive transition toward distributed mobile edge networks with virtualized network functions (VNFs) [e.g., firewalls, intrusion detection systems (IDSs), and transcoders] within the proximity of users. This transition will enable network services, especially Internet-of-Things (IoT) applications, to be provisioned as network slices with sequences of VNFs, in order to guarantee the performance and security of their continuous data and control flows. In this article, we study the problems of delay-aware network slicing for multicasting traffic of IoT applications in edge networks. We first propose exact solutions by formulating the problems into integer linear programs (ILPs). We further devise an approximation algorithm with an approximation ratio for the problem of delay-aware network slicing for a single multicast slice, with the objective to minimize the implementation cost of the network slice subject to its delay requirement constraint. Given multiple multicast slicing requests, we also propose an efficient heuristic that admits as many user requests as possible, through exploring the impact of a nontrivial interplay of the total computing resource demand and delay requirements. We then investigate the problem of delay-oriented network slicing with given levels of delay guarantees, considering that different types of IoT applications have different levels of delay requirements, for which we propose an efficient heuristic based on reinforcement learning (RL). We finally evaluate the performance of the proposed algorithms through both simulations and implementations in a real testbed. The experimental results demonstrate that the proposed algorithms are promising.

As a key technology of the Internet of Underwater Things (IoUT), underwater acoustic sensor networks (UASNs) have attracted considerable attention from both academia and industry. Due to specific characteristics of UWSNs, such as high latency, high mobility, and limited bandwidth, it is a challenge to design routing protocols for 3-D UASNs. In order to address these challenges, here, we propose a game-theoretic routing protocol (GTRP) for 3-D UASNs. First, the GTRP defines a forwarding area making the nodes closer to the destination inclined to forward. Then, it estimates the node degree in the forwarding area without broadcasting prior message periodically. Third, GTRP regards the forwarding process as a game. The number of participants in the game is the node degree information in the forwarding area, instead of the number of actual neighbors. To test the effectiveness of the proposed GTRP, we implement it and evaluate its performance in the Aqua-sim. The extensive simulations results indicate that GTRP significantly outperforms various existing protocols used for comparison, in terms of the number of received packets, the packet delivery fraction, and the end-to-end delay.

With the rapid development of the Internet-of-Things (IoT) networks, millions of IoT services provided through wireless networks are waiting for people’s exploration. Such a large number of heterogeneous IoT services produce huge amounts of data in almost real time, known asbig data, many of which cannot be measured or quantified. Hence, a recommended system that aims to deal with the unquantifiable big data is urgently needed. To solve the problem, we propose a novel quantile contextual tree-based multiarmed bandits algorithm to support the large-scale recommendation with both quantifiable and unquantifiable data. Furthermore, the high failure rate of communication has a serious influence on the recommendation accuracy of our system with the widely used D2D technology in today’s IoT network. To improve recommendation accuracy under the D2D communication, we take into account the feedback of historical service receivers and the historical successful delivery rate (SDP) of data transmission at the same time for the service recommendation system. We give theoretical analysis to prove a sublinear bound of the regret. Numerical experiments with tremendously large data sets show that we can balance the regret with the system time cost and guarantee a high SDP.

Faster-than-Nyquist signaling (FTNS) is capable of improving the spectral efficiency (SE) of communication systems. However, for conventional single-carrier FTNS (SC-FTNS) in which only symbol interval is reduced, the increase of SE is very limited due to the presence of inter-symbol interference (ISI) introduced by the FTNS. To deal with this problem, this paper proposes a new time-frequency compressed SC-FTNS (TFC-SC-FTNS) scheme that includes the conventional FTNS as a special case, to improve the SE via two dimensions simultaneously: time dimension by stacking symbols closer; frequency dimension by precoding to make the FTN signal spectrum more compact. Further, an optimization subject to a spectral mask constraint is performed on the precoder to suppress the ISI, according to a mean-square-error criterion, but the optimization problem is non-convex. A nontrivial contribution in the new scheme is that the non-convex problem is transformed into a convex one by a change of variable and an addition of admissibility constraint. Simulation results demonstrate that the proposed scheme significantly outperforms the conventional FTNS in terms of achievable SE or, equivalently, reception performance at a given SE. Further, with larger constellations applied, the gains of the TFC-FTNS increase.

Due to the increasing medical data for coronary heart disease (CHD) diagnosis, how to assist doctors to make proper clinical diagnosis has attracted considerable attention. However, it faces many challenges, including personalized diagnosis, high dimensional datasets, clinical privacy concerns and insufficient computing resources. To handle these issues, we propose a novel blockchain-enabled contextual online learning model under local differential privacy for CHD diagnosis in mobile edge computing. Various edge nodes in the network can collaborate with each other to achieve information sharing, which guarantees that CHD diagnosis is suitable and reliable. To support the dynamically increasing dataset, we adopt a top-down tree structure to contain medical records which is partitioned adaptively. Furthermore, we consider patients' contexts (e.g., lifestyle, medical history records, and physical features) to provide more accurate diagnosis. Besides, to protect the privacy of patients and medical transactions without any trusted third party, we utilize the local differential privacy with randomised response mechanism and ensure blockchain-enabled information-sharing authentication under multi-party computation. Based on the theoretical analysis, we confirm that we provide real-time and precious CHD diagnosis for patients with sublinear regret, and achieve efficient privacy protection. The experimental results validate that our algorithm {outperforms} other algorithm benchmarks on running time, error rate and diagnosis accuracy.

Stringent delay requirements of many mobile applications have led to the development of mobile edge clouds, to offer low latency network services at the network edges. Most conventional network services are implemented via hardware-based network functions, including firewalls and load balancers, to guarantee service security and performance. However, implementing hardware-based network functions usually incurs both a high capital expenditure (CAPEX) and operating expenditure (OPEX). Network Function Virtualization (NFV) exhibits a potential to reduce CAPEX and OPEX significantly, by deploying software-based network functions in virtual machines (VMs) on edge-clouds. We consider a fundamental problem of NFV-enabled multicasting in a mobile edge cloud, where each multicast request has both service function chain and end-to-end delay requirements. Specifically, each multicast request requires chaining of a sequence of network functions (referred to as a service function chain) from a source to a set of destinations within specified end-to-end delay requirements. We devise an approximation algorithm with a provable approximation ratio for a single multicast request admission if its delay requirement is negligible; otherwise, we propose an efficient heuristic. Furthermore, we also consider admissions of a given set of the delay-aware NFV-enabled multicast requests, for which we devise an efficient heuristic such that the system throughput is maximized, while the implementation cost of admitted requests is minimized. We finally evaluate the performance of the proposed algorithms in a real test-bed, and experimental results show that our algorithms outperform other similar approaches reported in literature.

The explosive growth of e-commerce has led to the development of the recommendation system. The recommendation system aims to provide a set of items that meet users’ personalized needs through analyzing users’ consumption records. However, the timeliness of purchasing data and the implicity of feedback data pose severe challenges for the existing recommendation methods. To alleviate these challenges, we exploit the user’s consumption records from the perspectives of user and item, by modeling the data on both item and user level, where the item-level value reflects the grade of item, and the user-level value reflects the user’s purchase intention. In this article, we collect the description information and the reviews of the items from public websites, then adopt sentiment analysis techniques to model the similarities on user level and item level, respectively. In particular, we extend the traditional latent factor model and propose two novel methods— I tem L evel Similarity M atrix F actorization (ILMF) and U ser L evel Similarity M atrix F actorization (ULMF)—by introducing two novel similarity measure methods. In ILMF and ULMF, the consistency between latent factors and explicit aspects is naturally incorporated into learning latent factors of the users and items, such that we can predict the users’ preferences on different items more accurately. Moreover, we propose I tem- U ser L evel Similarity M atrix F actorization (IULMF), which combines these two methods to study their contributions on the final performance. Experimental evaluations on the real datasets show that our methods outperform the baseline approaches in terms of both the precision and NDCG.

The consistency of a response to a given post at semantic-level and emotional-level is essential for a dialogue system to deliver human-like interactions. However, this challenge is not well addressed in the literature, since most of the approaches neglect the emotional information conveyed by a post while generating responses. This article addresses this problem by proposing a unified end-to-end neural architecture, which is capable of simultaneously encoding the semantics and the emotions in a post for generating more intelligent responses with appropriately expressed emotions. Extensive experiments on real-world data demonstrate that the proposed method outperforms the state-of-the-art methods in terms of both content coherence and emotion appropriateness.

Logistic regression (LR) is the most commonly used loss function in multi-label image classification. However, it suffers from class imbalance problem caused by the huge difference in quantity between positive and negative samples as well as between different classes. First, we find that feeding randomly generated noise samples into an LR classifier is an effective way to detect class imbalances, and further define an informative imbalance metric named inference tendency based on noise sample analysis. Second, we design an efficient moving average based method for calculating inference tendency, which can be easily done during training with negligible overhead. Third, two novel rectification methods called extremum shift (ES) and tendency constraint (TC) are designed to offset or constrain inference tendency in the loss function, and mitigate class imbalances significantly. Finally, comparative experiments with Resnet on Microsoft COCO, NUS-WIDE and DeepFashion demonstrate the effectiveness of inference tendency and the superiority of our approach over the baseline LR and several state-of-the-art alternatives.

We address the problem of semi-supervised video object segmentation (VOS), where the masks of objects of interests are given in the first frame of an input video. To deal with challenging cases where objects are occluded or missing, previous work relies on greedy data association strategies that make decisions for each frame individually. In this paper, we propose a novel approach to defer the decision making for a target object in each frame, until a global view can be established with the entire video being taken into consideration. Our approach is in the same spirit as Multiple Hypotheses Tracking (MHT) methods, making several critical adaptations for the VOS problem. We employ the bounding box (bbox) hypothesis for tracking tree formation, and the multiple hypotheses are spawned by propagating the preceding bbox into the detected bbox proposals within a gated region starting from the initial object mask in the first frame. The gated region is determined by a gating scheme which takes into account a more comprehensive motion model rather than the simple Kalman filtering model in traditional MHT. To further design more customized algorithms tailored for VOS, we develop a novel mask propagation score instead of the appearance similarity score that could be brittle due to large deformations. The mask propagation score, together with the motion score, determines the affinity between the hypotheses during tree pruning. Finally, a novel mask merging strategy is employed to handle mask conflicts between objects. Extensive experiments on challenging datasets demonstrate the effectiveness of the proposed method, especially in the case of object missing.

With the rapidly growing number of connected smart devices deployed and diverse services provided in the Internet of Things (IoT), selecting proper services for users is becoming more and more important. However, challenges exist as a result of the highly heterogeneous environments, characteristics of various kinds of users and the myriad services offered by many service providers, which have promising applications in the IoT era. In the meantime, users' contexts (e.g., location, time, and surroundings) are wildly utilized in the IoT scenario to better satisfy individuals' demands, raising privacy issues among people. To address these problems, we proposed a differentially private tree-based contextual online learning approach for IoT service selection to select suitable services for users. Leveraging on the historical records of services' and users' feedback, our algorithm achieves high prediction accuracy. Besides, instead of considering the services as individual items, we utilize a top-down cover tree structure to select services, which supports increasing large-scale dataset and diverse natural conditions. We theoretically prove that the accumulative regret of our approach has a sublinear bound and our experiment confirms that it can handle big data problems while achieving a balance between privacy-preserving level and service selection accuracy.

Faster-than-Nyquist signaling (FTNs) is capable of improving the signaling rate of communication system, while yielding the inter-symbol interference (ISI) complicating the receiver design. However, due to the unavoidable detection performance degradation when compression factor τ drops considerably below the Mazo limit, the achievable gain is limited. In this paper, preceding the FTN modulation, a precoding based data spreading is utilized to introduce an artificial interference, which aims to support a smaller τ that corresponds to achieving a higher capacity, at the cost of detection complexity. Further, we optimize the precoder by minimizing the mean square error (MSE) of the equalizer's output. Meanwhile, the problem is transformed and reformulated as a non-convex quadratically constrained and quadratic programming with one constraint (QCQP-1), where the consensus-alternating directions method of multipliers (ADMM) algorithm is utilized to iteratively pursue the solution. Simulation results justify the proposed scheme, where the capacity of 64-, 128-, and even 256-QAM Nyquist signaling can be achieved by precoding the 16-QAM FTNs, even without SNR loss at bit error rate (BER) of 10-5.

With the development of e-commerce and smart cities, utilizing unmanned vehicles to deliver packages has emerged as one of the most important methods to make customers receive packages efficiently and effectively. Hence, how to reasonably utilize multiple unmanned vehicles at the same time is a problem. Another main challenging issue is how to satisfy customers' personalized need. In this paper, we propose a novel context-aware multi-armed bandit-based online learning algorithm with active partition method for context space. To solve the massive injecting data flow problem, we utilize a tree-based structure expanding from top to bottom to choose different vehicles, which supports ever-increasing big metering datasets with historical and contextual information. We prove that our proposed context-aware online learning algorithm achieves sublinear regret performance. Experiment results show our proposal can enhance customers' satisfaction and reduce space cost tremendously.

Driven by stringent delay requirements of mobile applications, the mobile edge cloud has emerged as a major platform to offer low latency network services from the edge of networks. Most conventional network services are implemented via hardware-based network functions, such as firewalls and load balancers, to guarantee service security and performance. However, implementing such hardware-based network functions incurs high purchase and maintenance costs. Network function virtualization (NFV) as a promising technology exhibits great potential to reduce the purchase and maintenance costs by implementing network functions as software in virtual machines (VMs). In this paper, we consider a fundamental problem of NFV-enabled multicasting in a mobile edge cloud, where each multicast request requires to process its traffic in a specified sequence of network functions (referred to as a service chain) before the traffic from a source to a set of destinations. We devise a provable approximation algorithm with an approximation ratio for the problem if requests do not have delay requirements; otherwise, we propose an efficient heuristic for it. We also evaluate the performance of the proposed algorithms against the state-of-the-art NFV-enabled multicasting algorithms, and results show that our algorithms outperform their counterparts.

With the rapidly growing number of connected smart devices deployed and diverse services provided in the Internet of Things (IoT), service selection system, which discovers appropriate services for users, is becoming more and more important. However, challenges exist as a result of the highly heterogeneous environments, characteristics of various kinds of users and the expanding number of services offered by many service providers, which have promising applications in IoT. In the meantime, users' contexts (e.g., location, time, and surroundings), wildly utilized in the IoT scenario to better satisfy individuals' demands, raises privacy issues. To address these problems, we propose a privacy-preserving mobile edge computing (MEC) enabled context-aware service selection system over software defined networks (SDN) in IoT to provide suitable services for end users, where the edge nodes (ENs) cache and process information, acting as cooperative learners. Utilizing the users' feedback and the historical records monitored by the SDN-based network, our contextual online learning algorithm achieves high prediction accuracy. Besides, instead of considering them as individual items, we utilize a top-down cover tree structure to handle service data, which supports ever-increasing large-scale datasets and complex situations. We theoretically prove that the accumulative regret of our algorithm has a sublinear bound and our numerical results confirm that our algorithm can handle big data problems while achieving a balance between privacy-preserving level and service selection accuracy.

The enabling Internet-of-Things (IoT) technology has inspired a large number of novel platforms and applications. One popular IoT platform is unmanned aerial vehicles (UAVs, also known as drone). Benefiting from the intrinsic flexibility, convenience, and low cost, UAVs have great potentials to be utilized in various civil applications, including parcel delivery. However, suffering from limited payloads and battery capacities, it is uneconomical for UAVs to perform parcel delivery tasks independently. To conquer the drawbacks of low payloads and battery capacities, people propose to employ both trucks and drones to construct truck-drone parcel delivery systems. However, previous works only leverage either independent drones or truck-carried drones to collaborate with trucks. In contrast, in this article we propose to simultaneously employ trucks, truck-carried drones, and independent drones to construct a more efficient truck-drone parcel delivery system. We claim that such a hybrid parcel delivery system can fully exploit the complementary benefits of the three platforms. We propose a novel routing and scheduling algorithm, referred to as hybrid truck-drone delivery (HTDD) algorithm, to solve the hybrid parcel delivery problem, wherein M drones carried by M trucks, together with N independent drones, cooperate to deliver parcels to customers distributed in a wide region. The experimental results show that our algorithm outperforms the existing solutions which employ either independent drones or truck-carried drones.

We propose a novel context-aware task allocation framework for mobile crowdsensing in the scenario of edge computing to enable the crowdsensing platform effectively and real-timely handle large-scale crowdsensing tasks in smart city. The task allocation performs in both cloud computing layer and edge computing layer. It aims to combine the merits of cloud and edge computing, i.e., diminishing communication latency while guaranteeing overall scheduling. The cloud layer evaluates the participants' task-oriented reputation based on the participants' background information, task context, and historical feedbacks (i.e., rewards) and sends the edge layer the most promising subset of participants. Then the edge layer communicates with the participants for the real-time information and makes optimization based on the task requirement (e.g., maximizing the sensing coverage under the constraint of the task budget). In the cloud layer, we propose a privacy-preserving and contextual online learning algorithm to manage the participants' reputation. The algorithm can adapt the decision-making strategy based on previous performances of participants. In the edge layer, plenty of existing centralized task allocation strategies can be directly applied to optimize based on the participants' real-time information. Theoretical analysis shows that our proposal achieves sublinear regret and differential privacy for both requesters and participants. Experiments results validate that our proposed algorithm supports increasing big dataset while striking a balance between the privacy-preserving level and the prediction accuracy.

Energy-saving (ES) systems developed on the basis of the Internet-of-Things (IoT) by heavily relying on automated understanding of human behaviors and activities recognition is of paramount importance in smart home. However, classic approaches are incapable to understand the relations among users' contexts and ES of appliances very well, and they cannot handle massive metering and time-varying user context datasets. Moreover, privacy concern is thoroughly aroused from both the residential and utility provider sides as to its essentiality. To tackle these problems, we propose a privacy-preserving and residential context-aware online ES (PRCOES) system in an IoT-enabled smart home environment. We model the repeated interaction of ES of appliances and the activity recognition of user context as a contextual multiarmed bandits (CMAB) problem, where the context-aware online learning algorithm can predict appropriate energy offers (EOs) that could meet the users' satisfaction, task completion rate, and ES purposes for appliances. We utilize a tree-based structure expanding from top to bottom to recommend EOs, which supports ever-increasing big metering datasets with user context-awareness. Theoretical analysis shows that our proposal achieves sublinear regret and differential privacy for both residents and utility provider. Experiments results validate that PRCOES could enhance users' experience and prolong users' engagement in everyday ES while guarantee the privacy for both residents and utility provider.

Computerized healthcare has undergone rapid development thanks to the advances in medical imaging and machine learning technologies. Especially, recent progress on deep learning opens a new era for multimedia based clinical decision support. In this paper, we use deep learning with brain network and clinical relevant text information to make early diagnosis of Alzheimer's Disease (AD). The clinical relevant text information includes age, gender, and ApoE gene of the subject. The brain network is constructed by computing the functional connectivity of brain regions using resting-state functional magnetic resonance imaging (R-fMRI) data. A targeted autoencoder network is built to distinguish normal aging from mild cognitive impairment, an early stage of AD. The proposed method reveals discriminative brain network features effectively and provides a reliable classifier for AD detection. Compared to traditional classifiers based on R-fMRI time series data, about 31.21 percent improvement of the prediction accuracy is achieved by the proposed deep learning method, and the standard deviation reduces by 51.23 percent in the best case that means our prediction model is more stable and reliable compared to the traditional methods. Our work excavates deep learning's advantages of classifying high-dimensional multimedia data in medical services, and could help predict and prevent AD at an early stage.

Online video sharing (e.g., via YouTube or YouKu) has emerged as one of the most important services in the current Internet, where billions of videos on the cloud are awaiting exploration. Hence, a personalized video retrieval system is needed to help users find interesting videos from big data content. Two of the main challenges are to process the increasing amount of video big data and resolve the accompanying “cold start” issue efficiently. Another challenge is to satisfy the users’ need for personalized retrieval results, of which the accuracy is unknown. In this paper, we formulate the personalized video big data retrieval problem as an interaction between the user and the system via a stochastic process, not just a similarity matching, accuracy (feedback) model of the retrieval; introduce users’ real-time context into the retrieval system; and propose a general framework for this problem. By using a novelcontextualmultiarmed bandit-based algorithm to balance the accuracy and efficiency, we propose a context-based online big-data-oriented personalized video retrieval system. This system can support datasets that are dynamically increasing in size and has the property of cross-modal retrieval. Our approach provides accurate retrieval results withsublinearregret andlinearstorage complexity and significantly improves the learning speed. Furthermore, by learning for a cluster of similar contexts simultaneously, we can realize sublinear storage complexity with the same regret but slightly poorer performance on the “cold start” issue compared to the previous approach. We validate our theoretical results experimentally on a tremendously large dataset; the results demonstrate that the proposed algorithms outperform existing bandit-based online learning methods in terms of accuracy and efficiency and the adaptation from the bandit framework offers additional benefits.

The explosive growth of multimedia contents (MCs) in today's mobile social networks has pushed edge computing to face severe security and online big data-processing problems. On the one hand, the edge nodes (ENs) should help mobile users find, cache, and share MCs in the presence of an ever-increasing scale of multimedia big data. On the other hand, how to provide secure MC retrieval schemes to excludedishonest-and-maliciousuntrusted ENs and to prevent privacy breaches fromhonest-but-curiousENs and users is a challenging issue. To tackle these problems, we study the privacy-preserving and trustworthy MCs retrieval system to make personalized MC recommendations from ENs to users with big data support. In our framework, each EN is modeled as a distributed context-aware online learner. ENs collaborate to learn users’ preferences based on their contexts and previous behaviors and social intimacy. To support big data analytics, we establish an MC-cluster tree from top to the bottom to handle the dynamically varying cached MC datasets. A differentially private algorithm is proposed to preserve the data privacy among honest-but-curious ENs and users. To guarantee trustworthy edge computing, a trust evaluation mechanism is designed to evaluate the reliability of ENs. We further consider the structure of edge networks to improve the performance of our algorithm. Experimental results validate that our new framework can support increasing multimedia big datasets while striking a balance among privacy-preserving level, Trustworthy level, and caching MC prediction accuracy.

We consider the online shortest path routing (SPR) of a network with stochastically time varying link states under potential adversarial attacks. Due to the denial of service (DoS) attacks, the distributions of link states could be stochastic (benign) or adversarial at different temporal and spatial locations. Without any a priori, designing an adaptive and optimal DoS-proof SPR protocol to thwart all possible adversarial attacks is a very challenging issue. In this paper, we present the first such integral solution based on the multi-armed bandit (MAB) theory, where jamming is the adversarial strategy. By introducing a novel control parameter into the exploration phase for each link, a martingale inequality is applied in our formulated combinatorial adversarial MAB framework. The proposed algorithm could automatically detect the specific jammed and un-jammed links within a unified framework. As a result, the adaptive online SPR strategies with near-optimal learning performance in all possible regimes are obtained. Moreover, we propose the accelerated algorithms by multi-path route probing and cooperative learning among multiple sources, and study their implementation issues. Comparing to existing works, our algorithm has the respective 30.3% and 87.1% improvements of network delay for oblivious jamming and adaptive jamming given a typical learning period and a 81.5% improvement of learning duration under a specified network delay on average, while it enjoys almost the same performance without jamming. Lastly, the accelerated algorithms can achieve a maximal of 150.2% improvement in network delay and a 431.3% improvement in learning duration.

As a crucial task in heterogeneous distributed systems, DAG-scheduling models a scheduling application with a set of distributed tasks by a Direct Acyclic Graph (DAG). The goal is to assign tasks to different processors so that the whole application can finish as soon as possible. Task Duplication-Based (TDB) scheme is an important technique addressing this problem. The main idea is to duplicate tasks on multiple machines so that the results of the duplicated tasks are available on multiple machines to trade computation time for communication time. Existing TDB algorithms enumerate and test all possible duplication candidates, and only keep the candidates that can improve the overall scheduling. We observe that while a duplication candidate is ineffective at the moment, after other duplications have been applied, this ineffective duplication candidate can become effective, which in turn can cause other ineffective duplications to become effective. We call this phenomenon the chain reaction of task duplication. We propose a novel Task Duplication based Clustering Algorithm (TDCA) to improve the schedule performance by utilizing duplication task more thoroughly. TDCA improves parameter calculation, task duplication, and task merging. The analysis and experiments are based on randomly generated graphs with various characteristics, including DAG depth and width, communication-computing cost ration, and variant computation power of processors. Our results demonstrate that the TDCA algorithm is very competitive. It improves the schedule makespan of task duplication-based algorithms for heterogeneous systems for various communication-computing cost ratios.

We revisit the opportunistic scheduling problem in which a server opportunistically serves multiple classes of users under time-varying multi-state Markovian channels. The aim of the server is to find an optimal policy minimizing the average waiting cost of those users. Mathematically, the problem can be recast to a restless multiarmed bandit one, and a pivot to solve restless bandit by the Whittle index approach is to establish indexability. Despite the theoretical and practical importance of the Whittle index policy, the indexability is still open for opportunistic scheduling in the heterogeneous multi-state channel case. To fill this gap, we mathematically identify a set of sufficient conditions on a channel state transition matrix under which the indexability is guaranteed and consequently, the Whittle index policy is feasible. Furthermore, we obtain the closed-form Whittle index by exploiting the structural property of the channel state transition matrix. For a generic channel state transition matrix, we propose an eigenvalue-arithmetic-mean scheme to obtain the corresponding approximate matrix which satisfies the sufficient conditions, and consequently can get an approximate Whittle index. This paper constitutes a small step toward solving the opportunistic scheduling problem in its generic form involving multi-state Markovian channels and multi-class users.

Mobile Edge Computing (MEC) pushes computing functionalities away from the centralized cloud to the network edge, thereby meeting the latency requirements of many emerging mobile applications and saving backhaul network bandwidth. Although many existing works have studied computation of-floading policies, service caching is an equally, if not more important, design topic of MEC, yet receives much less attention. Service caching refers to caching application services and their related databases/libraries in the edge server (e.g. MEC-enabled BS), thereby enabling corresponding computation tasks to be executed. Because only a small number of application services can be cached in resource-limited edge server at the same time, which services to cache has to be judiciously decided to maximize the edge computing performance. In this paper, we investigate the extremely compelling but much less studied problem of dynamic service caching in MEC-enabled dense cellular networks. We propose an efficient online algorithm, called OREO, which jointly optimizes dynamic service caching and task offloading to address a number of key challenges in MEC systems, including service heterogeneity, unknown system dynamics, spatial demand coupling and decentralized coordination. Our algorithm is developed based on Lyapunov optimization and Gibbs sampling, works online without requiring future information, and achieves provable close-to-optimal performance. Simulation results show that our algorithm can effectively reduce computation latency for end users while keeping energy consumption low.

To adapt to the increasing storage demands and varying storage redundancy requirements, practical distributed storage systems need to support storage scaling by relocating currently stored data to different storage nodes. However, the scaling process inevitably transfers substantial data traffic over the network. Thus, minimizing the bandwidth cost of the scaling process is critical in distributed settings. In this paper, we show that optimal storage scaling is achievable in erasure-coded distributed storage based on network coding, by allowing storage nodes to send encoded data during scaling. We formally prove the information-theoretically minimum scaling bandwidth. Based on our theoretical findings, we also build a distributed storage system prototype NCScale, which realizes network-coding-based scaling while preserving the necessary properties for practical deployment. Experiments on Amazon EC2 show that the scaling time can be reduced by up to 50% over the state-of-the-art.

Data mining has heralded the major breakthrough in data analysis, serving as a “super cruncher” to discover hidden information and valuable knowledge in big data systems. For many applications, the collection of big data usually involves various parties who are interested in pooling their private data sets together to jointly train machine-learning models that yield more accurate prediction results. However, data owners may not be willing to disclose their own data due to privacy concerns, making it imperative to provide privacy guarantee in collaborative data mining over distributed data sets. In this paper, we focus on tree-based data mining. To begin with, we design novel privacy-preserving schemes for two most common tasks: regression and binary classification, where individual data owners can perform training locally in a differentially private manner. Then, for the first time, we design and implement a privacy-preserving system for gradient boosting decision tree (GBDT), where different regression trees trained by multiple data owners can be securely aggregated into an ensemble. We conduct extensive experiments to evaluate the performance of our system on multiple real-world data sets. The results demonstrate that our system can provide a strong privacy protection for individual data owners while maintaining the prediction accuracy of the original trained model.

It is critical to support efficient scaling in distributed storage systems so as to meet increasing storage demands with new storage nodes. However, the scaling process incurs substantial scaling bandwidth due to reorganizing currently stored data to new storage nodes. Recent work has applied network coding to minimize scaling bandwidth for a special case where (n, k) MDS codes are scaled to (n', k') MDS codes for n'-k'=n-k. In this paper, we extend the results and prove the minimum scaling bandwidth for a more general setting where (n, k) MDS codes are scaled to (n', k') MDS codes for and k'≥ k. Furthermore, we present a family of MDS code construction that achieves optimal scaling from (n, k) to (n', k') where k=k'.

In this paper, we investigate the usage of the convolutional neural network (CNN) to propose a novel end-to-end language processing structure to model textual data for this task. In particular, we propose a 3D CNN structure for the task, which is featured by spatial pyramid pooling (SPP). To our knowledge, it is the first time that 3D convolution and SPP structure are applied together in language processing issues. Compared with methods of 2D CNNs, the proposed method can effectively and efficiently capture the complicated internal relations in sentences. Furthermore, in previous work, the issue of sentence length variety is usually addressed by padding zero to make all sentences vectors to a fixed length, which causes too much redundant and useless noise. Inspired by the SPP structure for object detection in image processing, this issue can be well handled with the SPP, which divides the sentences into several length sections for respective pooling processing. Experiments are conducted for the task of sentence classification as well as relation classification. Experiments on Stanford Treebank, TREC, subj, and Yelp datasets demonstrate that our proposed method can outperform other state-of-the-art models, with respect to classification accuracy. Auxiliary attempts to leverage our method to SemEval-2010 Task 8 dataset further substantiate the model's capability of extracting features efficiently.

Cooperative multicast has been demonstrated to achieve significant performance gain over the classic source-destination transmission paradigm by exploiting spatial diversity through the participation of multiple relay nodes. As a major technical challenge, the selection of relays for a multicast session has significant impact on the multicast performance. The challenge is even more pronounced when the number of channels is limited as the relay selection is in this context coupled with channel allocation. The goal of this paper is to design a fair multicast relay selection scheme with limited channel resources. Specifically, we establish an analytical framework for this joint relay selection and channel allocation problem and develop a lexicographic max-min multicast relay selection scheme. Our design consists of two technical steps. First, we consider the maximization of the minimal data rate. By decoupling relay selection and channel allocation, the problem is transformed to a max-min-max problem, which is difficult to solve. To make this problem tractable, we reformulate it as a convex optimization problem via relaxation and smoothing, and prove the asymptotic equivalence from a geometrical perspective. Second, we propose an adjustment algorithm based on the initial max-min solution, and prove that the proposed scheme achieves lexicographic optimality. Finally, our proposed algorithm is evaluated by simulation to show its superiority over the conventional schemes.

Industrial fog computing deploys various industrial services, such as automatic monitoring/control and imminent failure detection, at the fog nodes (FNs) to improve the performance of industrial systems. Much effort has been made in the literature on the design of fog network architecture and computation offloading. This paper studies an equally important but much less investigated problem of service hosting where FNs are adaptively configured to host services for sensor nodes (SNs), thereby enabling corresponding tasks to be executed by the FNs. The problem of service hosting emerges because of the limited computational and storage resources at FNs, which limit the number of different types of services that can be hosted by an FN at the same time. Considering the variability of service demand in both temporal and spatial dimensions, when, where, and which services to host have to be judiciously decided to maximize the utility of the fog computing network. Our proposed fog configuration strategies are tailored to battery-powered FNs. The limited battery capacity of FNs creates a long-term energy budget constraint that significantly complicates the fog configuration problem as it introduces temporal coupling of decision making across the timeline. To address all these challenges, we propose an online distributed algorithm, called adaptive fog configuration (AFC), based on Lyapunov optimization and parallel Gibbs sampling. AFC jointly optimizes service hosting and task admission decisions, requiring only currently available system information while guaranteeing close-to-optimal performance compared to an oracle algorithm with full future information.

Distributed storage systems can store data with erasure coding to maintain data availability with low storage redundancy. One class of erasure coding is based on regenerating codes, which provably minimize the amount of data transferred for failure repair and realize the optimal tradeoff between the storage redundancy and the amount of traffic transferred for repair. Typical regenerating codes often require surviving storage nodes to encode their stored data for repair. In this paper, we study a framework called proxy-assisted regeneration, which offloads the repair process to a centralized proxy. We extend the previous applied work on proxy-assisted regeneration by providing theoretical validation. Specifically, we study a special class of regenerating codes called proxy-assisted minimum storage regenerating (PMSR) codes, which enable uncoded repair without the need of encoding in surviving nodes, while preserving the minimum storage redundancy and minimum amount of traffic transferred for repair. We formally prove the existence of PMSR codes for two configurations: 1) repairing single-node failures under double fault tolerance and 2) repairing double-node failures under triple fault tolerance. We also provide a semideterministic PMSR code construction for repairing single-node failures under double fault tolerance.

In the big data era, the generation of data presents some new characteristics, including wide distribution, high velocity, high dimensionality, and privacy concern. To address these challenges for big data analytics, we develop a privacy-preserving distributed online learning framework on the data collected from distributed data sources. Specifically, each node (i.e., data source) has the capacity of learning a model from its local dataset, and exchanges intermediate parameters with a random part of their own neighboring (logically connected) nodes. Hence, the topology of the communications in our distributed computing framework is unfixed in practice. As online learning always performs on the sensitive data, we introduce the notion of differential privacy (DP) into our distributed online learning algorithm (DOLA) to protect the data privacy during the learning, which prevents an adversary from inferring any significant sensitive information. Our model is of general value for big data analytics in the distributed setting, because it can provide rigorous and scalable privacy proof and have much less computational complexity when compared to classic schemes, e.g., secure multiparty computation (SMC). To tackle high-dimensional incoming data entries, we study a sparse version of the DOLA with novel DP techniques to save the computing resources and improve the utility. Furthermore, we present two modified private DOLAs to meet the need of practical applications. One is to convert the DOLA to distributed stochastic optimization in an offline setting, the other is to use the mini-batches approach to reduce the amount of the perturbation noise and improve the utility. We conduct experiments on real datasets in a configured distributed platform. Numerical experiment results validate the feasibility of our private DOLAs.

Nowadays, machine learning is becoming a new paradigm for mining hidden knowledge in big data. The collection and manipulation of big data not only create considerable values, but also raise serious privacy concerns. To protect the huge amount of potentially sensitive data, a straightforward approach is to encrypt data with specialized cryptographic tools. However, it is challenging to utilize or operate on encrypted data, especially to perform machine learning algorithms. In this paper, we investigate the problem of training high quality word vectors over large-scale encrypted data (from distributed data owners) with the privacy-preserving collaborative neural network learning algorithms. We leverage and also design a suite of arithmetic primitives (e.g., multiplication, fixed-point representation, sigmoid function computation, etc.) on encrypted data, served as components of our construction. We theoretically analyze the security and efficiency of our proposed construction, and conduct extensive experiments on representative real-world datasets to verify its practicality and effectiveness.

Shared edge computing platforms deployed at the radio access network are expected to significantly improve the quality-of-service delivered by application service providers (ASPs) in a flexible and economic way. However, placing edge service in every possible edge site by an ASP is practically infeasible due to the ASP’s prohibitive budget requirement. In this paper, we investigate the edge service placement problem of an ASP under a limited budget, where the ASP dynamically rents computing/storage resources in edge sites to host its applications in close proximity to end users. Since the benefit of placing edge service in a specific site is usually unknown to the ASPa priori, optimal placement decisions must be made while learning this benefit. We pose this problem as a novel combinatorial contextual bandit learning problem. It is “combinatorial” because only a limited number of edge sites can be rented to provide the edge service given the ASP’s budget. It is “contextual” because we utilize user context information to enable finer-grained learning and decision-making. To solve this problem and optimize the edge computing performance, we propose SEEN, a Spatial-temporal Edge sErvice placemeNt algorithm. Furthermore, SEEN is extended to scenarios with overlapping service coverage by incorporating a disjunctively constrained knapsack problem. In both cases, we prove that our algorithm achieves a sublinear regret bound when it is compared with an Oracle algorithm that knows the exact benefit information. Simulations are carried out on a real-world dataset, whose results show that SEEN significantly outperforms benchmark solutions.

Cooperative caching is a promising technology for enhancing user experience and reducing redundant transmissions through the participation of multiple caching nodes. In this paper, we design a clustering algorithm for the sectionalized caching, in which each user divides its caching space into two parts and the contents cached in these two parts are determined according to the individual interest and the joint interest of all users in the same cluster respectively. Different to most of the existing works forming the clusters based on the interest similarity of all files, we adopt the piecewise interest similarity as the criterion of clustering, which takes advantage of the content diversity and contributes to the reduction of the transmission delay. We measure the gain of the cooperation between two users and obtain the piecewise interest similarity for two users accordingly. Since the gain of clustered caching highly depends on the formed cluster structure, we estimate the gain of the clustered caching based on the piecewise interest similarities by online learning and propose an affinity propagation (AP) based clustering algorithm. Finally, our proposed clustering algorithm is evaluated by simulation to show its superiority over the conventional clustering algorithms.

To make every user conveniently have access to his or her most interested jobs and candidates (recommendation items) in the current employment market, the recruitment networks need to meet the demand of fast and accurate recommendation. But a key challenge is that the total items may have a large quantity in the big data scenarios. And another problem is that the personalization of different users is diverse. In order to handle these challenges, this paper proposes a mining and prediction system for job and candidate recommendation with contextual online learning. It predicts a proper item by utilizing the feedback reward of previous users in the nearby context region. Besides that, we introduce a Monte-Carlo Tree Search (MCTS) method in which the similar items can be amalgamated into a cluster to reduce the computing load. Our algorithm can achieve sublinear regret and space complexity. Finally, some experiments are conducted to test our algorithm based on a large database from \emph{Work4} (the global leader in social and mobile recruiting), which can show the outstanding performance of our algorithm when compared with other existing algorithms.

Orthogonal frequency division multiplexing (OFDM) can not meet the diverse scenarios of the future fifth generation (5G) networks due to its high out-of-band emission (OOBE), relatively low spectrum efficiency, and poor flexibility. In this paper, a nonuniform subband superposed OFDM (NSS-OFDM) scheme, based on a variable granularity (VG) spectrum allocation technique, is proposed as a candidate waveform for 5G. The VG method is exploited to divide the transmission band into a certain amount of subbands, each of which is applied to a specified application scenario through configuring the signaling parameters. To reduce the computational complexity, a multistage polyphase subfiltering architecture is utilized. Additionally, with subtly designed filters, the OOBE is significantly suppressed, which can minimize the frequency guard intervals between subbands. At the receiver, bit-error- rate is investigated for the additive white Gaussian noise and frequency selective channel. Simulation results show that the spectral efficiency, in terms of spectrum utilization rate, is increased up to 98.85% when ignoring the BER performance loss.

The state of art protocols in neighbor discovery in wireless sensor networks can be divided into two categories: the quorum based, and the co-primality based protocols, which are confirmed by simulations to have a small "theoretical" latency according to their wake-up schedule designs. However, these works fail to consider the collision of neighbor discovery beacons, when multiple pairs of nodes search for discovery simultaneously, which could lead to a greater "practical" latency. In this paper, we seek to reduce such practical latency in neighbor discovery protocols caused by the collision of neighbor discovery beacons by a dynamic slot-length control mechanism that allows each node to autonomously vary its slot length based on the detection of beacon collisions. We develop a real-world testbed of neighbor discovery on Android smartphones by implementing a few classic neighbor discovery protocols in literature, and evaluate their performance with and without the proposed slot-length control. Experimental results show that the slot-length control can help reduce the practical latency of neighbor discovery by more than 15% under both types of neighbor discovery protocols.

Person Re-Identification (person re-id) is a crucial task as its applications in visual surveillance and human-computer interaction. In this work, we present a novel joint Spatial and Temporal Attention Pooling Network (ASTPN) for video-based person re-identification, which enables the feature extractor to be aware of the current input video sequences, in a way that interdependency from the matching items can directly influence the computation of each other's representation. Specifically, the spatial pooling layer is able to select regions from each frame, while the attention temporal pooling performed can select informative frames over the sequence, both pooling guided by the information from distance matching. Experiments are conduced on the iLIDS-VID, PRID-2011 and MARS datasets and the results demonstrate that this approach outperforms existing state-of-art methods. We also analyze how the joint pooling in both dimensions can boost the person re-id performance more effectively than using either of them separately 1.

Thanks to the rapid development of information technology, the size of a wireless network is becoming larger and larger, which makes spectrum resources more precious than ever before. To improve the efficiency of spectrum utilization, game theory has been applied to study efficient spectrum sharing for a long time. However, the scale of wireless networks in existing studies is relatively small. In this paper, we introduce a novel game called aggregative game and use it to model spectrum sharing in large-scale, heterogeneous, and dynamic networks. Meanwhile, the massive usage of the spectrum leads to easier divulgence of privacy of spectrum users, which calls for privacy and truthfulness guarantees. In a large decentralized scenario, each user has no priori about other users’ channel access decisions, which forms an incomplete information game. A “weak mediator,” e.g., the base station or licensed spectrum regulator, is introduced and it turns the incomplete spectrum sharing game into a complete one. This is essential in reaching a Nash equilibrium (NE). By utilizing past channel access experience, we propose an online learning algorithm to improve the utility of each user. We show that the learning algorithm achieves an NE over time and provides no regret guarantee for each user. Specifically, our mechanism admits an approximateex-postNE, and is joint differentially private and incentive-compatible. Efficiency of the approximate NE is evaluated, and innovative scaling law results are disclosed. We also provide simulation results to verify our analysis.

This paper studies the problem of jamming-resistant spectrum aggregation and access (SAA) for energy-efficiency (EE) cognitive radio communications. We consider various jamming behaviors, where jammers may attack all available channels with arbitrarily changing strategies over time, attack a subset of the channels at certain time slots, or have different intelligence, i.e., oblivious or adaptive adversary, and so on. Without any priori knowledge about the channels and jammers, it is very challenging to design an efficient and practical jamming-resistant SAA algorithm to reach the optimal EE goal. In this paper, we utilize the advanced martingale concentration inequalities in an multi-armed bandits-based online learning framework to facilitate the optimal detection of various jamming behaviors. We first define a novel EE model for discontiguous orthogonal frequency division multiplexing to facilitate scalable SAA over distributed spectrum pools in practice. Then, the jamming-resistant dynamic channel access problem is formulated as a regret minimization problem. Meanwhile, an online stochastic gradient descent with bandit feedback procedure is adopted to allocate the transmit power. The proposed algorithm can autonomously detect the environmental features and find a near-optimal solution in each attacking scenario. Our algorithm is implemented with low complexity and with multiple users under some practical jamming scenarios. Extensive numerical studies show that under some practical jamming scenarios, our algorithm has an EE improvement of 45.3% over a fixed learning period, and an improvement of 82.5% in terms of learning duration compared with existing approaches.

Repair performance in hierarchical data centers is often bottlenecked by cross-rack network transfer. Recent theoretical results show that the cross-rack repair traffic can be minimized through repair layering, whose idea is to partition a repair operation into inner-rack and cross-rack layers. However, how repair layering should be implemented and deployed in practice remains an open issue. In this article, we address this issue by proposing a practical repair layering framework called DoubleR . We design two families of practical double regenerating codes (DRC), which not only minimize the cross-rack repair traffic but also have several practical properties that improve state-of-the-art regenerating codes. We implement and deploy DoubleR atop the Hadoop Distributed File System (HDFS) and show that DoubleR maintains the theoretical guarantees of DRC and improves the repair performance of regenerating codes in both node recovery and degraded read operations.

Modern e-health systems have undergone rapid development thanks to the advances in communications, computing and machine learning technology. Especially, deep learning has great superiority in image analysis and disease prediction. In this paper, we use Alzheimer's Disease (AD) as an example to show advantages of deep learning in diagnosing brain diseases and providing clinical decision support. Firstly, we convert raw functional magnetic resonance imaging (fMRI) to a matrix to represent activity of 90 brain regions. Secondly, to represent the functional connectivity between different brain regions, a correlation matrix is obtained by calculating the correlation between each pair of brain regions. In the next, a targeted autoencoder network is built to classify the correlation matrix, which is sensitive to AD. Finally, the experiment results show that our proposed method for AD prediction achieves much better effects than traditional means. It finds the correlations between different brain regions efficiently, provides strong reference for AD prediction. Compared to Support Vector Machine (SVM), about 25% improvement is gained in prediction accuracy. The e-health field becomes more complete and effective owing to that. Our work helps predict AD at an early stage and take measures to slow down or even prevent the onset of it.

WeChat is a mobile messaging application that has 549 million active users as of Q1 2015, and “WeChat Moments” (WM) serves its social-networking function that allows users to post/share links of web pages. WM differs from the other social networks as it imposes many restrictions on the information diffusion process to mitigate the information overload. In this paper, we conduct a measurement study on information diffusion in the WM network by crawling and analyzing the spreading statistics of more than 160,000 pages that involve approximately 40 million users. Specifically, we identify the relationship of the number of posted pages and the number of views, the diffusion path length, the similarity and distribution of users' locations as well as their connections with the GDP of the users' province. For each individual WM page, we measure its temporal characteristics (e.g., the life time, the popularity within a time period); for each individual user, we evaluate how many of, or how likely, one's friends will view his posted pages. Our results will help the business to decide when and how to release the marketing pages over WM for better publicity.

With the rapid growth of Internet technologies, cloud computing and social networks have become ubiquitous. An increasing number of people participate in social networks and massive online social data are obtained. In order to exploit knowledge from copious amounts of data obtained and predict social behavior of users, we urge to realize data mining in social networks. Almost all online websites use cloud services to effectively process the large scale of social data, which are gathered from distributed data centers. These data are so large-scale, high-dimension and widely distributed that we propose a distributed sparse online algorithm to handle them. Additionally, privacy-protection is an important point in social networks. We should not compromise the privacy of individuals in networks, while these social data are being learned for data mining. Thus we also consider the privacy problem in this article. Our simulations shows that the appropriate sparsity of data would enhance the performance of our algorithm and the privacy-preserving method does not significantly hurt the performance of the proposed algorithm.

In this paper, we address an important problem in the wireless monitoring, i.e., how to choose channels with best (or worst) qualities timely and accurately. We consider both scenarios of one or more sniffers simultaneously monitoring multiple channels in the same area. Since the channel information is initially unknown to the sniffers, we shall adopt learning methods during the monitoring to predict the channel condition by a short time of observation. We formulate this problem as a novel branch of the classic multi-armed bandit (MAB) problem, named exploration bandit problem, to achieve a trade-off between monitoring time/resource budget and the channel selection accuracy. In the multiple sniffer cases, including partly-distributed (with limited communications) and fully-distributed (without any communications) scenarios, we take communication costs and interference costs into account, and analyze how these costs affect the accuracy of channel selection. Extensive simulations are conducted and the results show that the proposed algorithms could achieve higher channel selection accuracy than other exploration bandit approaches, hence it proves the advantages of the proposed algorithms.

We consider the shortest path routing (SPR) problem of a network with time varying link metrics in unknown environments. Due to potential denial of service attacks, the distributions of link states could be stochastic (benign or i.i.d.), contaminated or adversarial (non-i.i.d.) at different temporal and spatial locations. Without any a priori, designing an adaptive SPR protocol to cope with all possible situations in practice optimally is a very challenging issue. In this paper, we present the first solution by formulating it as a multi-armed bandit problem. By introducing novel control parameters to explore link conditions, our proposed algorithms could automatically detect features of the environment within a unified framework and find the optimal SPR strategies with almost optimal learning performance in all possible cases over time. Moreover, we study important issues related to the practical implementation, such as decoupling route selection with multi-path route probing, cooperative learning among multiple sources, the cold-start issue and delayed feedback of our algorithm. Nonetheless, the proposed SPR algorithms can be implemented with low complexity and they are proved to scale very well with the network size. The efficacy of the proposed solutions is verified by simulations from the real tracedriven datasets. Comparing to existing approaches in a typical network scenario, our algorithm has a 65.3 percent improvement of network delay given a learning period and a 81.5 percent improvement of learning duration under a specified network delay.

With the rapid growth in multimedia services and the enormous offers of video content in online social networks, users have difficulty in obtaining their interests. Therefore, various personalized recommendation systems have been proposed. However, they ignore that the accelerated proliferation of social media data has led to the big data era, which has greatly impeded the process of video recommendation. In addition, none of them has considered both the privacy of users' contexts (e.g., social status, ages, and hobbies) and video service vendors' repositories, which are extremely sensitive and of significant commercial value. To handle these problems, we propose a cloud-assisted differentially private video recommendation system based on distributed online learning. In our framework, service vendors are modeled as distributed cooperative learners, recommending videos according to user's context, while simultaneously adapting the video-selection strategy based on user-click feedback to maximize total user clicks (reward). Considering the sparsity and heterogeneity of big social media data, we also propose a novel geometric differentially private model, which can greatly reduce the performance loss. Our simulation shows the proposed algorithms outperform other existing methods and keep a delicate balance between the total reward and privacy preserving level.

Designing efficient channel access schemes for wireless communications without any prior knowledge about the nature of environments has been a very challenging issue, in which the channel state distribution of all spectrum resources could be entirely or partially stochastic or adversarial at different times and locations. In this paper, we propose an online learning algorithm for adaptive channel access of wireless communications in unknown environments based on the theory of multiarmed bandits (MAB) problems. By automatically tuning two control parameters, i.e., learning rate and exploration probability, our algorithms could find the optimal channel access strategies and achieve the almost optimal learning performance over time in different scenarios. The quantitative performance studies indicate the superior throughput gain when compared with previous solutions and the flexibility of our algorithm in practice, which is resilient to both oblivious and adaptive jamming attacks with different intelligence and attacking strength that ranges from no-attack to the full-attack of all spectrum resources. We conduct extensive simulations to validate our theoretical analysis.

In this paper, we consider joint end-to-end congestion control and multiuser scheduling in multi-hop wireless networks, where orthogonal frequency-division multiple access (OFDMA) is the air interface. Compared to its single-channel counterpart, this problem is significantly more challenging because subchannel assignment and power allocation for all nodes and network traffic need to be jointly optimized. To avoid cochannel inference, we first consider an OFDMA subchannel separation problem in multi-hop communication scenarios. Then, we address this cross-layer optimization problem by decomposing it into two subproblems: a TCP layer congestion control for traffic allocation and a weighted scheduling for OFDMA subchannel assignment and power control. Observing that the scheduling for multiuser OFDMA in multi-hop wireless networks is inherently NP-hard, we propose a low-complexity algorithm that ensures a solution converging to the optimum. The performance of the algorithm is demonstrated through simulations.

Cooperative transmission technique is now widely considered as a promising approach to combat fading and achieve energy efficiency in wireless networks. In this paper we focus on the routing problem in energy-constrained wireless sensor networks (WSNs), of which a cooperative MISO-based routing strategy is adopted. We first analyze the physical layer energy consumption model of cooperative transmission in the scenarios of one hop and hop-to-hop for energy-efficient routing in order to prolong the network lifetime. Based on this analysis, we disclose how the energy-efficient network routing problem is tightly related to the inter-cluster MISO node and hop-to-hop relay node selection. As we noticed, the problem of energy-efficient cooperative routing is NP-hard innately which is difficult to implement in a totally distributive approach. Due to these analysis, a feasible algorithm with minimum cost is thus proposed. In the simulation part, we prove that our protocol can prolong the network lifetime tremendously when choose appropriate transmission parameters. Moreover, as an example, we simulate a typical network scenario which indicates our protocol is more energy efficient when comparing with vMIMO scheme in our previous work.

In cognitive radio networks, power control is necessary to not only decrease the interference among the secondary users (SUs), but also avoid negative impact to the primary users (PUs). Prevalent research works on power control are mainly focus on maximizing SINR as the QoS requirement of SUs under the interference power constraint for PUs. We note that besides achieving a high SINR to guarantee reliable data transmissions, SUs also require to support heterogenous services with different transmission rates. In order to provide flexible transmission rates to each SU, efficient use of networks radio resource requires transmission rate control in addition to transmit power control. In this paper, we consider the problem of joint power and rate control for SUs in cognitive radio network by using non-cooperative game theory. We study how to jointly allocate optimal transmit power and transmission rate given certain QoS requirement of SUs. We analysis of existence, uniqueness and Pareto efficiency of Nash equilibrium for our game. The performance of our proposed joint power and rate control algorithm is investigated by numeral results.

Cooperative MIMO transmission technique is considered as one of the effective solutions to reduce the energy consumption in wireless sensor networks. However, the existing cooperative MIMO based protocols only focused on how to reduce the energy consumption, but for some applications such as audio/video surveillance sensor network, a large amount of gathered data (formulated as network utility function) and long network lifetime are both required. In those cases, the two performance parameters should be considered and jointly optimized during the protocol design. In this paper, we first model and analyze the energy consumption and channel capacity of Multi-hop cooperative MIMO transmission. Then, we propose a joint network lifetime and utility optimization model based on NUM (network utility maximization) approach. Finally, we use the dual decomposition technique to solve the primal optimization problem and get a distributed algorithm. Simulation results show that, by using our distributed algorithm, the lifetime and utility of cooperative MIMO sensor network can converge to Pareto optimal trade-off.

Next Generation wireless networks are evolving towards all-data system which are expected to support a variety of application services with diverse transmission rates. Meanwhile, since most of the mobile terminals in wireless networks are battery-powered, to use energy efficiently, each terminal needs to transmit just enough power to achieve the desired transmission rate without causing excessive interference in the network. In this paper, a game-theoretic framework is used to study the joint power and rate control problem on the energy efficiency of wireless data network. A energy-efficient non-cooperative joint power and rate control game is thus introduced in which each user seeks to choose its possible transmit power and transmission rate in order to maximize its own utility while satisfying its target SINR as quality-of service (QoS) requirement. The utility function here we adopt is especially suitable for energy-constrained networks. We introduce pricing of transmit power into the utility function which not only improves the overall system performance, but also obtains Pareto Improvement when compared to the game with no pricing. The existence, uniqueness, best-response strategies and Pareto efficiency of Nash Equilibrium for the proposed game are proved. Based on these analysis, we present a distributive joint power and rate control algorithm. In the simulation part, we investigate the best pricing factor and compare our proposed algorithm with alternative algorithms developed by using game theory.

We present a cluster-based virtual multiple-input single-output (vMISO) multihop cooperative transmission protocol in wireless sensor networks. In order to construct the cooperative transmission, we arrange each hop into two consecutive time slots: the intra-cluster slot, that accounts for data sharing within the cluster, and the inter-cluster slot for transmission between clusters. In this protocol, we investigate a novel optimal cluster formation method and best cooperative transmission routing selection base on the network energy consumption status. Also, an optimization model is developed to find the optimum number of cooperative nodes, clusters and transmission rates. Simulation results show that the protocol can save energy significantly.

As Wireless Sensor Networks find more and more application in daily life, energy efficiency of these networks, which require unattended operations, is of paramount concern. In this paper, we propose an energy-efficient virtual MIMO transmission scheme based on V-BLAST technique. The proposed scheme does not require transmitter-side cooperation (joint STBC encoding), which is more suitable for the energy-constrained wireless sensor networks compared with the Alamouti-based virtual MIMO scheme. Simulation results show significant energy savings and network lifetime extension can be achieved by using our scheme.