The emergence of deploying Deep neural network (DNN) services on edge servers has spurred research into efficiently provisioning inference services. However, previous studies have neglected to consider the implications of different types of DNN and varying quality of service (QoS) requirements on QoS violation rates. In this article, we propose a novel framework, named Coinf, for scheduling heterogeneous DNN inference tasks on edge servers. Coinf has the following four advantages to effectively handle attribute analysis, performance balancing, parallel execution, and model accuracy: (1) It enables efficient profiling of domain-specific attributes of various DNN tasks during the offline stage, achieved by constructing a regression model to predict the end-to-end latency of each task. (2) By utilizing the predicted execution time, Coinf achieves a commendable balance among inference latency, system throughput, and QoS violation rate. (3) It employs emerging deep reinforcement learning (DRL) to aggregate individual DNN tasks into batches, enabling concurrent parallel execution. (4) Coinf preserves the accuracies of the provided DNN models by not modifying them. Numerical experiments are constructed to validate the reliability and efficiency of Coinf in handling heterogeneous inference tasks.

Federated Learning (FL) enables collaborative model training while preserving data privacy, but the presence of noisy labels in local datasets remains a significant challenge, particularly under heterogeneous noise conditions and class imbalance. In this work, we introduce a novel Multi-Stage Robust Federated Learning (MRFL) framework to address these issues. In the warm-up noise detection stage, MRFL computes per-class average losses on each client and employs a Gaussian mixture model to accurately identify clients with substantial label noise. In the subsequent noise-robust training stage, a robust loss function and noise solver are designed to distinguish clean from noisy samples, while semi-supervised learning is used to recover valuable information from tail classes. Moreover, a robust weighted aggregation strategy is adopted to mitigate the adverse effects of noisy clients. Extensive experiments on CIFAR-10/100-LT and ICH datasets demonstrate that MRFL outperforms state-of-the-art methods in federated noisy label learning scenarios characterized by data heterogeneity and imbalance.

Recent years have seen extensive demonstration of the validity and reliability of emotional information contained within electroencephalography (EEG) data. Nonetheless, challenges persist in the realm of cross-subject emotion recognition utilizing EEG data. Most existing research methods focus on intra-subject emotion recognition, while their application effectiveness in cross-subject emotion recognition is relatively inferior. Therefore, we propose a network based on a bidirectional selective state space model (SSM) with multi-scale convolution and additive gated attention. Specifically, the model initially captures global emotion-related information from differential entropy data using a bidirectional selective SSM, while simultaneously extracting local information at various scales through the multi-scale convolutional module. Subsequently, the model extracts deep emotion-related information from the data through the additive gated attention and ultimately inputs the processed data into a multilayer perceptron (MLP) to obtain emotion classification results. Experimental results validate the model's efficacy in cross-subject emotion recognition tasks, achieving accuracy rates of 86.22% and 74.98% on the SEED and SEED-IV datasets, respectively. By leveraging the attention mechanism, the study explored the differential contributions of various cortical areas to emotional processing, providing insights into the neural mechanisms underlying emotional responses.

Recent years have seen extensive demonstration of the validity and reliability of emotional information contained within electroencephalography (EEG) data. Nonetheless, challenges persist in the realm of cross-subject emotion recognition utilizing EEG data. Most existing research methods focus on intra-subject emotion recognition, while their application effectiveness in cross-subject emotion recognition is relatively inferior. Therefore, we propose a network based on a bidirectional selective state space model (SSM) with multi-scale convolution and additive gated attention. Specifically, the model initially captures global emotion-related information from differential entropy data using a bidirectional selective SSM, while simultaneously extracting local information at various scales through the multi-scale convolutional module. Subsequently, the model extracts deep emotion-related information from the data through the additive gated attention and ultimately inputs the processed data into a multilayer perceptron (MLP) to obtain emotion classification results. Experimental results validate the model’s efficacy in cross-subject emotion recognition tasks, achieving accuracy rates of 86.22% and 74.98% on the SEED and SEED-IV datasets, respectively. By leveraging the attention mechanism, the study explored the differential contributions of various cortical areas to emotional processing, providing insights into the neural mechanisms underlying emotional responses.

In vehicular networks enhanced by uncrewed aerial vehicles (UAVs), vehicle state information is efficiently collected, and traffic safety is assured. UAVs, serving as aerial base stations, enable vehicle network access and provide edge computing services in the absence of roadside units (RSUs). This study explores a multi-UAV-assisted vehicular network, where multiple UAVs collaboratively offer services to vehicles. The goal is to minimize task completion time by optimizing trajectory planning, spectrum resource allocation, and dynamic data offloading. An enhanced multiagent deep deterministic policy gradient (MADDPG) algorithm is introduced to address the optimization challenge in cooperative multi-UAV scenarios. Within this framework, each UAV, acting as an agent, devises strategies for movement, data offloading, and resource allocation based on the current states of vehicles and fellow UAVs. The simulation results reveal that the proposed algorithm improves task completion efficiency and ensures vehicle Quality of Service (QoS) over existing benchmarks.

Integrating data security with resource management enhances security, efficiency, and reliability of blockchain-enabled mobile edge computing (MEC) systems. However, challenges such as secure data storage, timely task execution, and limited parallelism introduce complexities in task offloading decisions and resource allocation strategies. To address these challenges, the task latency minimization problem in blockchain-enabled MEC networks is formulated as an NP-hard optimization problem. The model incorporates constraints on parallelism, partial task offloading, bandwidth and computation resource allocation among mobile users (MUs) and edge servers (ESs). To enhance the reliability and transparency of data storage, a dual-blockchain framework is proposed, consisting of multiple MU blockchains and a dedicated ES blockchain. To tackle the NP-hard problem, the original optimization problem is decomposed into multiple sub-problems, facilitating parameter decoupling. An alternating optimization algorithm is employed to refine task offloading decisions and resource allocation of MUs and ESs with limited parallelism. The ESs update their strategies iteratively based on feedback mechanisms. Additionally, a task prioritization formulation is developed to enhance scalability, considering sub-level task importance, urgency, and first-level task classification. Extensive simulation experiments demonstrate that the proposed algorithm achieves lower task latency compared to existing methods across varying network sizes, offloading schemes, and parallelism constraints. By optimizing the parallel processing of tasks, the waiting latency of this algorithm is reduced on average by 35. 35%, 57. 16% and 35. 35% compared to other methods, respectively.

Vehicular edge computing (VEC) has emerged as a promising paradigm for efficient processing of computation-intensive and delay-sensitive tasks by coordinating service placement and task offloading. Existing research mainly focused on edge-edge and edge-cloud collaborations to enhance system performance and resource utilization. However, the potential of vehicle-vehicle collaboration remains under-explored. To bridge this gap, we propose a novel three-layer VEC architecture integrating vertical collaboration across different layers with horizontal collaboration within the same layer (vehicle-edge-cloud collaboration). Recognizing the dynamic nature of the Internet of Vehicles, we introduce link duration constraints to quantify the impact of vehicles’ mobility on wireless communications. We formulate a mixed-integer nonlinear programming problem for joint service placement, task offloading, and computing resource allocation to minimize the total task completion delay of vehicles. To solve it, a two-stage heuristic algorithm is designed, including a semidefinite relaxation-based approximation method for the task offloading and computing resource allocation problem without storage capacity constraints and a heuristic approach for the service placement problem. Extensive simulations conducted on synthetic and realistic road topologies demonstrate that the proposed algorithm can obtain feasible solutions and achieve significantly lower delay than five benchmark methods.

The emergence of new applications has led to a high demand for mobile-edge computing (MEC), which is a promising paradigm with a cloud-like architecture deployed at the network edge to provide computation and storage services to mobile users (MUs). Since MEC servers have limited resources compared to the remote cloud, it is crucial to optimize resource allocation in MEC systems and balance the load among cooperating MEC servers. Caching application data for different types of computing services (CSs) at MEC servers can also be highly beneficial. In this article, we investigate the problem of hierarchical joint caching and resource allocation in a cooperative MEC system, which is formulated as an infinite-horizon cost minimization Markov decision process (MDP). To deal with the large state and action spaces, we decompose the problem into two coupled subproblems and develop a hierarchical reinforcement learning (HRL)-based solution. The lower layer uses the deep$Q$network (DQN) to obtain service caching and workload offloading decisions, while the upper layer leverages DQN to obtain load balancing decisions among cooperative MEC servers. The feasibility and effectiveness of our proposed schemes are validated by our evaluation results.

For Internet and cloud computing service providers, running massive geo-distributed data centers incurs prodigious electricity cost and water consumption as well as carbon emission rooted in electricity generation. Thus, it is critical significant for providers to lower down the operation cost of data centers. In this article, we investigate the problem of energy management for geo-distributed data centers with renewable resources and energy storages. We aim to minimize the long-term operation cost including electricity cost, water consumption, and carbon emission by leveraging the spatiotemporal diversity of these system states. To this end, we first formulate the cost minimization problem as a stochastic optimization problem, then we adopt the Lyapunov optimization technique to design a close-to-optimal online algorithm which only needs the current system information and achieves a delicate tradeoff between system cost and performance of delay tolerant workloads. To reduce the computational complexity and unnecessary communication, we further propose a distributed algorithm based on the distributed computing framework alternating direction method of multipliers (ADMM), which enables each data center to make their own control decisions. Based on the real-world traces and extensive simulations, we demonstrate the effectiveness of our proposed algorithms.

Mobile Edge Computing (MEC) brings abundant cloud resources to the edge of the network and provides great opportunities to improve user's quality of experience. While many recent studies have investigated the problem of computation offloading, service caching is also an important design topic of MEC. Service caching stores application-related databases or libraries in advance and enables corresponding user tasks to be offloaded. Due to the limited resources in the edge server, service caching decisions have to be made judiciously to maximize the system performance. In this paper, we study the problem of joint service caching, computation offloading, transmission and computing resource allocation in a general scenario of multiple users with multiple tasks. We aim to minimize the overall computation and delay costs for all users and formulate the optimization problem as a quadratically constrained quadratic program (QCQP) which is non-convex and NP-hard. To solve this challenging problem, we propose an efficiently approximate algorithm based on semidefinite relaxation (SDR) approach and alternating optimization which always computes a locally optimal solution. Moreover, we extend the study to the scenario where each user has a computation cost constraint. Simulation results show that our algorithm can minimize the system cost effectively by utilizing the available system resources.

Mobile-edge computing (MEC) has aroused significant attention for its performance to accelerate application's operation and enrich user's experience. With the increasing development of green computing, energy harvesting (EH) is considered as an available technology to capture energy from circumambient environment to supply extra energy for mobile devices. In this paper, we propose an online dynamic tasks assignment scheduling to investigate the tradeoff between energy consumption and execution delay for an MEC system with EH capability. We formulate it into an average weighted sum of energy consumption and execution delay minimization problem of mobile device with the stability of buffer queues and battery level as constraints. Based on the Lyapunov optimization method, we obtain the optimal scheduling about the CPU-cycle frequencies of mobile device and transmit power for data transmission. Besides, the dynamic online tasks offloading strategy is developed to modify the data backlogs of queues. The performance analysis shows the stability of the battery energy level and the tradeoff between energy consumption and execution delay. Moreover, the MEC system with EH devices and task buffers implements the high energy efficient and low latency communications. The performance of the proposed online algorithm is validated with extensive trace-driven simulations.