Synchronous Dataflow Graphs (SDFGs) are widely employed to model real-time embedded systems and streaming data processing, where throughput serves as a critical measure of computational efficiency. Parametric Shortest Path (PSP) algorithms offer an effective means of analyzing the optimal throughput of Homogeneous SDFGs (HSDFGs). However, applying PSP algorithms to general SDFGs typically requires a conversion to HSDFGs, which introduces additional overhead in graph transformation and may result in exponential growth in graph size. This paper proposes an extension to a traditional PSP algorithm, enabling direct throughput analysis of SDFGs without explicit conversion to HSDFGs. Furthermore, a graph size reduction technique is incorporated to further optimize the runtime of the proposed algorithm. Experimental results demonstrate that the proposed algorithm achieves, on average, a shorter runtime than three state-of-the-art algorithms. The advantage of the proposed algorithm scales with the size of the SDFG, achieving a speedup of up to 39.05x over the fastest of the three baseline algorithms.

Large language models (LLMs) are increasingly powering web-based applications, whose effectiveness relies on fine-tuning with large-scale instruction data. However, such data often contains valuable or sensitive information that limits its public sharing among business organizations. Federated learning (FL) enables collaborative fine-tuning of LLMs without accessing raw data. Existing approaches to federated LLM fine-tuning usually adopt a uniform model architecture, making it challenging to fit highly heterogeneous client-side data in varying domains and tasks, e.g., hospitals and financial institutions conducting federated fine-tuning may require different LLM architectures due to the distinct nature of their domains and tasks. To address this, we propose FedAMoLE, a lightweight personalized FL framework that enables data-driven heterogeneous model architectures. It features a heterogeneous mixture of low-rank adaptation (LoRA) experts module to aggregate architecturally heterogeneous models and a reverse selection-based expert assignment strategy to tailor model architectures for each client based on data distributions. Experiments across seven scenarios demonstrate that FedAMoLE improves client-side performance by an average of 5.97% over existing approaches while maintaining practical memory, communication, and computation overhead.

Knowledge graphs (KGs) have been widely adopted to mitigate data sparsity and address cold-start issues in recommender systems. While existing KGs-based recommendation methods can predict user preferences and demands, they fall short in generating explicit recommendation paths and lack explainability. As a step beyond the above methods, recent advancements utilize reinforcement learning (RL) to find suitable items for a given user via explainable recommendation paths. However, the performance of these solutions is still limited by the following two points. (1) Lack of ability to capture contextual dependencies from neighboring information. (2) The excessive reliance on short recommendation paths due to efficiency concerns. To surmount these challenges, we propose a category-aware dual-agent reinforcement learning (CADRL) model for explainable recommendations over KGs. Specifically, our model comprises two components: (1) a category-aware gated graph neural network that jointly captures context-aware item representations from neighboring entities and categories, and (2) a dual-agent RL framework where two agents efficiently traverse long paths to search for suitable items. Finally, experimental results show that CADRL outperforms state-of-the-art models in terms of both effectiveness and efficiency on large-scale datasets.

Synchronous Dataflow graphs (SDFGs) are widely used to model streaming applications that exhibit data-driven and iterative execution patterns. Graph conversion techniques such as retiming, unfolding, and pipelining are commonly used to optimize the iteration periods (IPs) of SDFGs. In this paper, we propose an extension of the graph conversion based pipelining approach for single-rate SDFGs to multi-rate SDFGs. A new perspective on pipelining is introduced, where the pipelining of a general-time SDFG can be viewed as the retiming of a unit-time SDFG. Based on this perspective, we prove that optimal pipelining can always achieve an IP less than 1 time unit longer than the optimal IP for an SDFG. Furthermore, an efficient optimal SDFG pipelining algorithm called GCP-SDFG is presented. Experimental results show that GCP-SDFG has significant advantages in IP minimizing and runtime relative to three state-of-the-art retiming or pipelining algorithms.

Recent advances of consensus control have made it significant in multi-agent systems such as in distributed machine learning, distributed multi-vehicle cooperative systems. However, during its application it is crucial to achieve resilience and privacy; specifically, when there are adversary/faulty nodes in a general topology structure, normal agents can also reach consensus while keeping their actual states unobserved. In this article, we modify the state-of-the-art Q-consensus algorithm by introducing predefined noise or well-designed cryptography to guarantee the privacy of each agent state. In the former case, we add specified noise on agent state before it is transmitted to the neighbors and then gradually decrease the value of noise so the exact agent state cannot be evaluated. In the latter one, the Paillier cryptosystem is applied for reconstructing reward function in two consecutive interactions between each pair of neighboring agents. Therefore, multi-agent privacy-preserving resilient consensus (MAPPRC) can be achieved in a general topology structure. Moreover, in the modified version, we reconstruct reward function and credibility function so both convergence rate and stability of the system are improved. The simulation results indicate the algorithms’ tolerance for constant and/or persistent faulty agents as well as their protection of privacy. Compared with the previous studies that consider both resilience and privacy-preserving requirements, the proposed algorithms in this article greatly relax the topological conditions. At the end of the article, to verify the effectiveness of the proposed algorithms, we conduct two sets of experiments, i.e., a smart-car hardware platform consisting of four vehicles and a distributed machine learning platform containing 10 workers and a server.

Distributed deep reinforcement learning(DDRL) has been used in distributed systems to better improve the adaptability. However, DDRL-based systems are also inevitably under the threat of Byzantine workers. There is an urgent need to enhance the resilience of the DDRL-based system against Byzantine failures. This paper proposes a resilient mechanism for mitigating the influence of Byzantine workers on DDRL-based systems. First, we formalize the DDRL-based system as a multi-armed bandit model for well capturing the collective effect of workers on the whole learning process, and then transforming the resilient mechanism design problem into the sampling policy optimization problem. Second, we propose a self-adaptation process for filtering out the harmful data generated by Byzantine workers and theoretically give a mathematical analysis of the understanding, demonstrating its effectiveness under ideal conditions. Third, based on a typical DDRL-based system (i.e., Asynchronous Advantage Actor-Critic, A3C), we implement a resilient distributed A3C (ReD-A3C). With extensive experiments on the DDRL benchmark tasks, we show that ReD-A3C outperforms available Byzantine tolerant approaches.

Recent advances in Multi-agent Reinforcement Learning (MARL) have made it possible to implement various tasks in cooperative as well as competitive scenarios through trial and error, and deep neural networks. These successes motivate us to bring the mechanism of MARL into the Multi-agent Resilient Consensus (MARC) problem that studies the consensus problem in a network of agents with faulty ones. Relying on the natural characteristics of the system goal, the key component in MARL, reward function, can thus be directly constructed via the relative distance among agents. Firstly, we apply Deep Deterministic Policy Gradient (DDPG) on each single agent to train and learn adjacent weights of neighboring agents in a distributed manner, that we call Distributed-DDPG (D-DDPG), so as to minimize the weights from suspicious agents and eliminate the corresponding influences. Secondly, to get rid of neural networks and their time-consuming training process, a Q-learning based algorithm, called Q-consensus, is further presented by building a proper reward function and a credibility function for each pair of neighboring agents so that the adjacent weights can update in an adaptive way. The experimental results indicate that both algorithms perform well with appearance of constant and/or random faulty agents, yet the Q-consensus algorithm outperforms the faulty ones running D-DDPG. Compared to the traditional resilient consensus strategies, e.g., Weighted-Mean-Subsequence-Reduced (W-MSR) or trustworthiness analysis, the proposed Q-consensus algorithm has greatly relaxed the topology requirements, as well as reduced the storage and computation loads. Finally, a smart-car hardware platform consisting of six vehicles is used to verify the effectiveness of the Q-consensus algorithm by achieving resilient velocity synchronization.

This article focuses on the research of a general time-varying nonlinear optimization (TVNO) problem solving especially in a noise-disturbance environment. For addressing this problem more efficiently, a new noise-enduring and finite-time convergent design formula is suggested to establish a novel zeroing neural network (NZNN). In contrast to the initial zeroing neural network or the noising-enduring zeroing neural network, which either only achieves finite-time convergence or only suppresses external disturbances, the merit of the proposed NZNN model is able to find an error-free optimal solution in a finite time under various different types of external noises. In addition, the detailed mathematical analyses about finite-time convergence and noise endurance are given to prove the excellent characteristics of the NZNN model. Numerical comparative results are provided to demonstrate the accuracy, efficiency, and advantages of the NZNN model for TVNO under various types of external disturbances. Robotic tracking example further validates the applicability of the NZNN model especially in a noise-disturbance environment.

Mobile crowdsourcing‐based applications, widely popular, exploit the sensing data crowdsourced from smartphone users without putting any burden on the extra cost of data sensing and collection. However, user participation in crowdsourcing incurs resource cost, such as battery, bandwidth, thus it is critical to design incentive mechanisms for propelling user's participation. Previous diverse incentive mechanisms designed for crowdsourcing applications only focus on users' contribution for reward allocation, while ignore another important property, i.e. fairness, users' reward should be corresponding with their cost. In this study, the authors first introduce a new concept called rate of return (RoR), defined as the ratio of received reward and incurred cost for each user, to demonstrate the property of fairness. With the goal of guarantee, the fairness of reward allocation for each user in a multiple‐task system, three algorithms, consensus‐based reward allocation, consensus‐based balanced topology reward allocation and Gossip‐based reward allocation are proposed for the demands of various scenarios, in which the RoR values are synchronised by optimising the fairness function in either centralised or decentralised manner. Through rigorous theoretical analysis and extensive simulations, it is finally demonstrated that the proposed reward allocation algorithms have the good property of fairness with quick convergence.

This paper presents a hierarchical structure to solve the consensus problem of multiagent systems. The new scheme divides the agents into several groups, with each group containing a value concerning all of the intragroup agents' states, which we call group information. For each single agent, it receives not only the agent information from its intragroup neighbors, but also the group information from its neighboring groups. It is then shown that global consensus can be achieved under the proposed scheme in both discrete time and continuous time. Moreover, a sufficient condition to achieve average consensus is provided. This hierarchical model can be well used in the PageRank algorithm to reduce the communication loads, and to reveal the attractors for Boolean networks by reducing the computational complexity.

This paper addresses distributed transmit beamforming problems based on a mobile assister node. Assuming that the distance to the receiver and the direction-of-destination (DoD) can be estimated, a stop-and-go strategy is proposed for a mobile assister node such that it moves gradually along the direction towards the receiver and provides SNR feedback information for transmitters. By receiving the SNR feedback information at each step, the transmitting sensors update their phases to maximize the SNR at the assister end. We then show that when the assister node reaches a threshold distance, the maximum SNR at the assister implies an approximate optimum SNR at the receiver end. Thus, the transmit beamforming problem is solved though the receiver initially is out of the communication range of the transmitters and is not able to provide SNR feedback information.

The paper studies the problem of distributed transmit beamforming with autonomous and self-organizing mobile antennas. The objective is to design a distributed algorithm for a network of autonomous mobile robots with carry-on antennas so that they can form a functional antenna array and cooperatively transmit messages to a remote station. Note that the spatial relationship of the antennas also contributes to the directionality of the reception or transmission of a signal. In the paper, by exploiting the mobility of the antennas, we show that optimal beamforming can be achieved by reconfiguring the spatial relationship of the mobile antennas in a completely distributed fashion. A probability-based coordination scheme utilizing only the signal-to-ratio (SNR) feedback from the receiver is presented to update the positions of the antennas ensuring that they eventually converge to a global optimal configuration maximizing the SNR at the receiver. It is noticed that the spatial configuration of the antennas can also address the phase synchronization issue in transmit beamforming.