Article
Engineering, Multidisciplinary
Ti Zhou, Man Lin
Summary: This paper introduces an intelligent Linux frequency Deep-Recurrent-Q-Network (DRQN) governor for dedicated applications with deadline requirements running on CPSS devices via machine learning. The governor can autonomously develop tradeoff policies to meet user needs with low overhead.
IEEE TRANSACTIONS ON NETWORK SCIENCE AND ENGINEERING
(2022)
Article
Engineering, Electrical & Electronic
Jangsaeng Kim, Young-Tak Seo, Wonjun Shin, Woo Young Choi, Byung-Gook Park, Jong-Ho Lee
Summary: In this study, we propose an efficient exploration method using the low-frequency noise of synaptic devices, which is applicable to hardware-based deep Q-networks. The proposed method achieves exploration efficiently with a relatively low hardware burden compared to other published studies. A rounded dual-channel flash memory cell is utilized as the synaptic device. The performance evaluation based on a simple Snake game demonstrates that the proposed system achieves similar performance to that using the epsilon-greedy exploration method, even with a low noise level of the synaptic devices and without the need for an additional circuit.
IEEE ELECTRON DEVICE LETTERS
(2023)
Article
Computer Science, Artificial Intelligence
Hanchi Huang, Li Shen, Deheng Ye, Wei Liu
Summary: A novel master-slave architecture is proposed to solve the top-K combinatorial multiarmed bandits problem with nonlinear bandit feedback and diversity constraints. It significantly outperforms existing algorithms in recommendation tasks.
IEEE TRANSACTIONS ON NEURAL NETWORKS AND LEARNING SYSTEMS
(2023)
Article
Computer Science, Artificial Intelligence
Francisco Munguia-Galeano, Ah-Hwee Tan, Ze Ji
Summary: This article proposes a new framework called IECR for learning from contextual information in order to solve complex computational problems. The framework represents each state using contextual key frames and extracts the affordances of the state using loss functions. By developing four new algorithms and evaluating them in five discrete environments, it is shown that all the algorithms that use contextual information significantly outperform the state-of-the-art methods.
IEEE TRANSACTIONS ON NEURAL NETWORKS AND LEARNING SYSTEMS
(2023)
Article
Robotics
Daniel Honerkamp, Tim Welschehold, Abhinav Valada
Summary: Despite its importance, mobile manipulation remains a significant challenge due to the need for integration of end-effector trajectory generation and navigation skills. Existing methods struggle with controlling the large configuration space and navigating dynamic and unknown environments. In this work, we introduce a new approach called Neural Navigation for Mobile Manipulation (NM2-M-2) that extends the decomposition of tasks in complex obstacle environments, enabling robots to perform a broader range of tasks in real-world settings. The approach demonstrates capabilities in extensive simulation and real-world experiments.
IEEE TRANSACTIONS ON ROBOTICS
(2023)
Article
Computer Science, Information Systems
Lihua Ai, Bin Tan, Jiadi Zhang, Rui Wang, Jun Wu
Summary: Mobile-edge computing (MEC) technology is used to provide computing resources for mobile devices and enhance the intelligence of the Internet of Things (IoT). This article proposes a hybrid optimization problem considering the time-varying channel and available computing resources of MEC servers. A reinforcement learning algorithm is used to predict current channel state information (CSI) and optimize task offloading, while convex optimization methods are employed for dynamic resource allocation. Simulation experiments demonstrate the effectiveness of the proposed algorithms.
IEEE INTERNET OF THINGS JOURNAL
(2023)
Article
Computer Science, Artificial Intelligence
Zhiwei Shang, Renxing Li, Chunhua Zheng, Huiyun Li, Yunduan Cui
Summary: In this article, a novel reinforcement learning approach, continuous dynamic policy programming (CDPP), is proposed to improve learning stability and sample efficiency in RL methods with continuous actions. The proposed method utilizes relative entropy regularization and Monte Carlo estimation to enhance the learning process, and outperforms baseline approaches in several tasks.
IEEE TRANSACTIONS ON NEURAL NETWORKS AND LEARNING SYSTEMS
(2023)
Article
Automation & Control Systems
Chenxi Wang, Youtian Du, Yuanlin Chang, Zihao Guo, Yanhao Huang
Summary: This article presents a human-machine collaborative framework for controlling line flow in power systems. Experimental results demonstrate that this approach can significantly improve the regulation performance.
IEEE TRANSACTIONS ON INDUSTRIAL INFORMATICS
(2023)
Article
Computer Science, Information Systems
Shengxiang Li, Ou Li, Guangyi Liu, Siyuan Ding, Yijie Bai
Summary: This paper introduces a novel policy gradient method to improve sample efficiency and reduce variance in training. Experimental results show that the proposed method can learn more steadily and achieve higher performance than existing methods.
Article
Computer Science, Information Systems
Peng Liu, Zhe Liu, Ji Wang, Zifu Wu, Peng Li, Huijuan Lu
Summary: This paper investigates the offloading scheduling issue for cyclical tasks and proposes the Multi-AGV Cyclical Offloading Optimization (MCOO) algorithm to reduce conflicts when multiple AGVs access the same ES. The algorithm divides the offloading optimization problem into two parts: finding the optimal allocation of tasks for a single AGV using load balancing and greedy algorithms under limited conditions, and asynchronously training multiple AGVs using Reinforcement Learning-based A3C algorithm to optimize the offloading scheme. Simulation results show that the MCOO algorithm improves global offloading performance in terms of task volume and adaptability compared to baseline algorithms.
JOURNAL OF CLOUD COMPUTING-ADVANCES SYSTEMS AND APPLICATIONS
(2022)
Article
Computer Science, Information Systems
Jihye Ryu, Juhyeok Kwon, Jeong-Dong Ryoo, Taesik Cheung, Jinoo Joung
Summary: In this study, a timeslot scheduling algorithm for traffic with similar requirements but different priorities is designed using a double deep q-network (DDQN), a reinforcement learning algorithm. The behavior of the DDQN agent is evaluated by defining a reward function based on the difference between estimated delay and packet deadline, as well as packet priority. The simulation shows that the designed algorithm outperforms existing algorithms in terms of more packets arrived within the deadline. The proposed DDQN-based scheduler can be implemented in upcoming frameworks for autonomous network scheduling.
Article
Computer Science, Information Systems
Xulong Li, Yunhui Qin, Jiahao Huo, Wei Huangfu
Summary: This article proposes a heuristic-assisted multi-agent reinforcement learning framework for joint optimization of computation offloading and resource allocation in mobile-edge computing. By introducing heuristic search and the soft actor-critic algorithm, the framework demonstrates superiority in the field.
IEEE INTERNET OF THINGS JOURNAL
(2023)
Article
Physics, Multidisciplinary
Zhimin He, Lvzhou Li, Shenggen Zheng, Yongyao Li, Haozhen Situ
Summary: The paper introduces a variational quantum compiling (VQC) algorithm based on reinforcement learning to automatically design the structure of quantum circuits without human intervention. The agent is trained to select quantum gates and the qubits they act on. Simulation results show that this method can reduce the number of quantum gates and decrease errors in quantum algorithms on NISQ devices.
NEW JOURNAL OF PHYSICS
(2021)
Article
Computer Science, Information Systems
Bolei Zhang, Bin Tang, Fu Xiao
Summary: This paper proposes a coordination method for mobile users to choose appropriate edge servers by offloading computation tasks in IoT applications. By introducing an additional virtual agent to broadcast public messages to mobile users, the users can make coordinated decisions.
IEEE INTERNET OF THINGS JOURNAL
(2023)
Article
Computer Science, Information Systems
Houchun Yin, Zhiwen Yu, Liang Wang, Jiangtao Wang, Lei Han, Bin Guo
Summary: This research addresses the issue of instant sensing and instant actuation (ISIA) in MCS, proposing the ISIATasker framework to optimize task allocation through clustering sensing locations, selecting sensors, and optimizing task distribution paths.
IEEE INTERNET OF THINGS JOURNAL
(2022)
Article
Computer Science, Theory & Methods
Weiguang Liu, Jinhua Cui, Tiantian Li, Junwei Liu, Laurence T. Yang
Summary: This paper presents MLCache, a space-efficient shared cache management scheme for NVMe SSDs. By learning the impact of reuse distance on cache allocation and building a workload-generic neural network model, MLCache achieves efficient space allocation decisions. Additionally, MLCache proposes an efficient parallel writing back strategy to improve fairness.
IEEE TRANSACTIONS ON PARALLEL AND DISTRIBUTED SYSTEMS
(2023)
Article
Computer Science, Artificial Intelligence
Jing Yang, Laurence T. Yang, Hao Wang, Yuan Gao, Yaliang Zhao, Xia Xie, Yan Lu
Summary: This paper investigates the progress and function of representation learning models adopted in knowledge fusion and reasoning, providing new perspectives and ideas for scholars. The paper comprehensively reviews classic methods and investigates advanced and emerging works. Additionally, an integrated knowledge representation learning framework and tensor-based knowledge fusion and reasoning models are proposed.
INFORMATION FUSION
(2023)
Article
Computer Science, Artificial Intelligence
Sushil Kumar Singh, Laurence T. Yang, Jong Hyuk Park
Summary: This article proposes a scheme called FusionFedBlock to address privacy issues in Industry 5.0 by combining blockchain and federated learning. In the scheme, industry departments can perform local learning updates and communicate with a global model, with validation conducted through a blockchain network. The scheme demonstrates excellent performance in privacy preservation and accuracy improvement.
INFORMATION FUSION
(2023)
Article
Computer Science, Hardware & Architecture
Shunli Zhang, Laurence T. Yang, Yue Zhang, Xiaokang Zhou, Zongmin Cui
Summary: This article presents a data-driven system-level design framework for responsible cyber-physical-social systems (CPSS), addressing ethical and legal concerns regarding decision-making in artificial intelligence systems.
Editorial Material
Computer Science, Hardware & Architecture
Sahil Garg, Jia Hu, Giancarlo Fortino, Laurence T. Yang, Mohsen Guizani, Xianjun Deng, Danda B. Rawat
Article
Engineering, Multidisciplinary
Yunzhi Xia, Xianjun Deng, Lingzhi Yi, Laurence T. Yang, Xiao Tang, Chenlu Zhu, Zhongping Tian
Summary: This paper proposes a 6G IoT coverage hole recovery algorithm based on Mobile Edge Computing (MEC) and Artificial Intelligence (AI). The algorithm utilizes the fusion model of the disc model and the confident information model to guide the movement of mobile edge nodes and repair the coverage holes through repeated games based on Q-learning.
IEEE TRANSACTIONS ON NETWORK SCIENCE AND ENGINEERING
(2023)
Article
Engineering, Electrical & Electronic
Xiaokang Zhou, Xuzhe Zheng, Xuesong Cui, Jiashuai Shi, Wei Liang, Zheng Yan, Laurence T. Yang, Shohei Shimizu, Kevin I-Kai Wang
Summary: This paper proposes a three-layer Federated Reinforcement Learning (FRL) framework with an end-edge-cloud structure, incorporating a digital twin system. It aims to enable lightweight model training and real-time processing in high-speed mobile networks. The proposed dual-reinforcement learning scheme and model splitting scheme effectively reduce communication costs and improve the non-IID problem.
IEEE JOURNAL ON SELECTED AREAS IN COMMUNICATIONS
(2023)
Article
Automation & Control Systems
Huazhong Liu, Jiawei Wang, Xiaoxue Yin, Jihong Ding, Laurence T. Yang, Tong Yao, Jing Yang, Yuan Gao
Summary: This article proposes a tensor-train (TT)-based multiuser multivariate multiorder (3M) physical Markov prediction approach for multimodal industrial trajectory pattern mining. The proposed approach improves the computational efficiency up to three times compared with the original tensor-based 3M approach, while ensuring basically consistent prediction accuracy.
IEEE TRANSACTIONS ON INDUSTRIAL INFORMATICS
(2023)
Article
Engineering, Civil
Zongmin Cui, Zhixing Lu, Laurence T. Yang, Jing Yu, Lianhua Chi, Yan Xiao, Shunli Zhang
Summary: There are three key roles in Intelligent Transportation Systems: driver, vehicle, and road. Existing static interactions among them are not dynamic enough, and unable to reflect changes in driver preferences, vehicle conditions, and road conditions. To address this issue, a data-driven Cloud-Fog-Edge Collaborative Driver-Vehicle-Road (CFEC-DVR) framework is proposed, which continuously adapts and evolves to provide better ITS services for humans.
IEEE TRANSACTIONS ON INTELLIGENT TRANSPORTATION SYSTEMS
(2023)
Article
Engineering, Civil
Jia Zhou, Guoqi Xie, Haibo Zeng, Weizhe Zhang, Laurence T. Yang, Mamoun Alazab, Renfa Li
Summary: In this paper, a clock-skew-based approach is proposed to pinpoint the sender and detect intrusion on proprietary CAN bus. By analyzing data from real vehicles, a box-plot algorithm based on score mechanism is presented to filter and describe the hardware characteristics of ECUs accurately.
IEEE TRANSACTIONS ON INTELLIGENT TRANSPORTATION SYSTEMS
(2023)
Article
Computer Science, Information Systems
Xu Li, Feilong Tang, Luoyi Fu, Jiadi Yu, Long Chen, Jiacheng Liu, Yanmin Zhu, Laurence T. Yang
Summary: The provisioning of satellite controllers has a significant impact on the performance of software-defined satellite networks. The challenge lies in achieving low control overhead throughout the operation period, despite the difficulty in predicting network load accurately. Existing methods struggle to address this issue, leading to frequent controller migrations. In this paper, we propose globally optimized strategies utilizing current network load information and introduce approximate and heuristic algorithms to solve the Controller Provisioning Problem in SDSNs.
IEEE TRANSACTIONS ON MOBILE COMPUTING
(2023)
Article
Computer Science, Information Systems
Hui Zhu, Xiaohu Tang, Laurence Tianruo Yang, Chao Fu, Shuangrong Peng
Summary: This paper proposes an efficient sampling-based scheme for collecting and analyzing key-value data. By utilizing probability sampling and optimizing budget allocation, the proposed scheme improves the probability of users submitting valid key-value data and achieves accurate frequency and mean estimation.
INFORMATION SCIENCES
(2023)
Article
Computer Science, Artificial Intelligence
Bocheng Ren, Laurence T. Yang, Qingchen Zhang, Jun Feng, Xin Nie
Summary: Various stream learning methods are emerging to provide solutions for artificial intelligence in streaming data scenarios. However, when each data stream is oriented to a different target space, it becomes impracticable to use the previous approaches. Therefore, we propose an adaptive learning scheme using tensor and meta-learning to mitigate domain shift and improve performance for few-shot streaming tasks.
IEEE TRANSACTIONS ON NEURAL NETWORKS AND LEARNING SYSTEMS
(2023)
Article
Computer Science, Artificial Intelligence
Xiaokang Wang, Lei Ren, Ruixue Yuan, Laurence T. Yang, M. Jamal Deen
Summary: In this article, a cloud-edge-aided quantized tensor-train distributed long short-term memory (QTT-DLSTM) method is presented as an approach for efficiently processing CPSS big data. By decomposing the multi-attributes CPSS big data into the QTT form, and utilizing a distributed cloud-edge computing model, the proposed method effectively improves training efficiency.
IEEE TRANSACTIONS ON NEURAL NETWORKS AND LEARNING SYSTEMS
(2023)
Article
Engineering, Multidisciplinary
Bocheng Ren, Laurence T. Yang, Qingchen Zhang, Jun Feng, Xin Nie
Summary: The development of artificial intelligence and the Internet of Things has provided opportunities for healthcare transformation, but data privacy and security remain concerns. We propose a blockchain-powered intelligent healthcare system that uses tensor meta-learning models to efficiently model heterogeneous healthcare data and protect private data.
IEEE TRANSACTIONS ON NETWORK SCIENCE AND ENGINEERING
(2023)