Article
Engineering, Multidisciplinary
Zakaria Abou El Houda, Abdelhakim Senhaji Hafid, Lyes Khoukhi, Bouziane Brik
Summary: Data-driven machine and deep learning can improve clinical decisions, but the lack of realistic and recent medical data hinders its adoption. This paper proposes HealthFed, a framework that uses federated learning and blockchain technologies to enable privacy-preserving and distributed learning among clinicians, ensuring the privacy of each participant and maintaining collaboration in a secure and flexible way.
IEEE TRANSACTIONS ON NETWORK SCIENCE AND ENGINEERING
(2023)
Article
Computer Science, Artificial Intelligence
Yuanyishu Tian, Yao Wan, Lingjuan Lyu, Dezhong Yao, Hai Jin, Lichao Sun
Summary: The fast growth of pre-trained models has revolutionized natural language processing, becoming the dominant technique for various NLP applications. However, pre-training requires significant training data and computing resources, making it impossible for individual clients to conduct. To enable clients with limited computing capability to participate in pre-training large models, FEDBERT proposes a federated learning approach that achieves excellent performance without sharing raw data.
ACM TRANSACTIONS ON INTELLIGENT SYSTEMS AND TECHNOLOGY
(2022)
Article
Engineering, Multidisciplinary
Zeou Hu, Kiarash Shaloudegi, Guojun Zhang, Yaoliang Yu
Summary: In this work, federated learning is formulated as multi-objective optimization and a new algorithm called FedMGDA+ is proposed, which guarantees fairness and robustness while maintaining individual performance for participating users.
IEEE TRANSACTIONS ON NETWORK SCIENCE AND ENGINEERING
(2022)
Article
Computer Science, Information Systems
Dinh C. Nguyen, Ming Ding, Quoc-Viet Pham, Pubudu N. Pathirana, Long Bao Le, Aruna Seneviratne, Jun Li, Dusit Niyato, H. Vincent Poor
Summary: The article discusses the concept of FLchain in MEC networks, focusing on privacy protection, security, cross-device collaboration, and resource allocation. FLchain integrates FL and blockchain technology, presenting a promising paradigm for intelligent MEC networks.
IEEE INTERNET OF THINGS JOURNAL
(2021)
Review
Computer Science, Theory & Methods
Hao Li, Chengcheng Li, Jian Wang, Aimin Yang, Zezhong Ma, Zunqian Zhang, Dianbo Hua
Summary: Artificial intelligence (AI) has contributed to the rapid development of healthcare, addressing complex medical problems. However, the lack of standardization in patient electronic medical records and legal and ethical requirements for patient information privacy hinders widespread AI integration. Federated learning, combined with privacy-preserving algorithms, can overcome data fragmentation and improve security and computational efficiency when combined with blockchain and edge computing. This paper reviews recent research on federated learning in healthcare, explores its architectures and classification models, and analyzes its advantages and security risks in medical applications. Standard privacy protection methods are introduced and the current state of federated learning and healthcare applications is discussed, concluding with a summary and future outlook.
FUTURE GENERATION COMPUTER SYSTEMS-THE INTERNATIONAL JOURNAL OF ESCIENCE
(2023)
Article
Computer Science, Information Systems
Shuai Yu, Xu Chen, Zhi Zhou, Xiaowen Gong, Di Wu
Summary: This paper discusses the challenges posed by smart cities, healthcare systems, and smart vehicles on the capability and connectivity of Internet of Things (IoT) devices, introducing concepts such as multiaccess edge computing and ultradense edge computing. It proposes an intelligent ultradense edge computing framework that integrates blockchain and artificial intelligence into 5G networks, and utilizes a two-timescale deep reinforcement learning approach to optimize computation offloading, resource allocation, and service caching placement, while leveraging federated learning for data privacy protection._simulation results demonstrate the effectiveness of the proposed algorithm in reducing task execution time by up to 31.87%.
IEEE INTERNET OF THINGS JOURNAL
(2021)
Article
Quantum Science & Technology
Amandeep Singh Bhatia, Sabre Kais, Muhammad Ashraful Alam
Summary: In recent years, the concept of federated machine learning has gained traction among scientists to address privacy concerns. The combination of machine learning and quantum computing is a disruptive force in various industries. Researchers have developed a hybrid quantum-classical algorithm called a quanvolutional neural network for efficient execution on quantum hardware. This study evaluates the performance of the algorithm on real-world data partitioned among healthcare institutions/clients, demonstrating its potential benefits in reducing communication rounds and maintaining accuracy.
QUANTUM SCIENCE AND TECHNOLOGY
(2023)
Article
Computer Science, Information Systems
Fan Zhang, Shaoyong Guo, Xuesong Qiu, Siya Xu, Feng Qi, Zhili Wang
Summary: With the rapid development of 5G and 6G technology, sharing cross-domain scattered data and enhancing data value transmission has become an inevitable trend. Federated learning (FL) is a new data-sharing technology that has gained wide attention due to its intelligence and privacy computing capabilities. However, FL lacks supervision in the application process, leading to unreliable calculation processes and result transmissions. To address this, we propose extending the computing and supervision capabilities of blockchain using state channels to create sandboxes and instantiate FL tasks for a trust supervision mechanism. Through theoretical analysis and experimental verification, our designed supervision mechanism demonstrates excellent performance in improving system security, resisting malicious attacks, and enhancing data model quality.
IEEE INTERNET OF THINGS JOURNAL
(2023)
Article
Computer Science, Theory & Methods
Youyang Qu, Md Palash Uddin, Chenquan Gan, Yong Xiang, Longxiang Gao, John Yearwood
Summary: Federated Learning (FL), driven by the growth of machine learning and Artificial Intelligence as well as emerging privacy concerns, has gained popularity in recent years. FL allows a central server and local end devices to maintain the same model by exchanging model updates instead of raw data, thus protecting the privacy of sensitive data. However, the performance of FL with a central server is limited, and new threats are emerging. To accelerate the adoption of FL, blockchain-enabled FL has attracted attention as it provides theories and techniques to enhance FL performance. This survey aims to comprehensively summarize and evaluate existing blockchain-enabled FL variants, identify emerging challenges, and propose potential research directions in this under-explored field.
ACM COMPUTING SURVEYS
(2023)
Article
Computer Science, Information Systems
Raushan Myrzashova, Saeed Hamood Alsamhi, Alexey V. Shvetsov, Ammar Hawbani, Xi Wei
Summary: Recently, innovations in the Internet of Medical Things (IoMT), information and communication technologies, and machine learning (ML) have enabled smart healthcare. Federated learning (FL) overcomes the challenges of centralized data storage and provides high-level security and privacy for smart healthcare. Combining blockchain and FL can further enhance the competency of healthcare by managing data in a decentralized manner and utilizing dispersed clinical data fully.
IEEE INTERNET OF THINGS JOURNAL
(2023)
Article
Computer Science, Information Systems
Hajar Moudoud, Soumaya Cherkaoui
Summary: The Metaverse is a shared virtual space where users can interact with each other and a virtual environment in real-time. Advances in virtual and augmented reality technologies have increased the popularity of the Metaverse, but also raise concerns about data privacy and security. Federated learning and blockchain-based solutions can address these issues and enhance security and trustworthiness in the Metaverse.
Article
Energy & Fuels
Siyuan Chen, Jun Zhang, Yuyang Bai, Peidong Xu, Tianlu Gao, Huaiguang Jiang, Wenzhong Gao, Xiang Li
Summary: This article introduces a Blockchain Enabled Intelligence of Federated Systems (BELIEFS) for cooperative control in multi-regional large-scale power systems using a multi-agents system. The system is capable of defending against malicious attacks and has been validated for effectiveness and efficiency through mathematical modeling and comparative experiments.
Article
Chemistry, Multidisciplinary
Remi Gosselin, Loic Vieu, Faiza Loukil, Alexandre Benoit
Summary: As privacy concerns rise, decentralized approaches like Federated Learning (FL) offer potential improvements in privacy protection and generalization behaviors. However, security issues such as poisoning and adversarial attacks pose threats to the model. This study comprehensively discusses the privacy and security issues in FL and identifies state-of-the-art approaches to address them.
APPLIED SCIENCES-BASEL
(2022)
Article
Computer Science, Information Systems
Marco Fisichella, Gianluca Lax, Antonia Russo
Summary: This paper introduces the application of federated learning in machine learning, which utilizes a privacy-preserving distributed framework to allow multiple client devices to participate in global model training and decide whether to share data.
INFORMATION SCIENCES
(2022)
Article
Computer Science, Artificial Intelligence
Taotao Wang, Soung Chang Liew, Shengli Zhang
Summary: This study applies reinforcement learning to derive an optimal Bitcoin-like blockchain mining strategy without requiring knowledge of the network model. By designing a new multidimensional RL algorithm, it achieves performance approaching the optimal mining strategy even in time-varying blockchain networks.
INTERNATIONAL JOURNAL OF INTELLIGENT SYSTEMS
(2021)
Article
Management
Dang-Trinh Nguyen, Dinh-Thien-Vuong Doan, Nguyen-Ngoc Cuong Tran, Duc-Hoc Tran
Summary: Project managers aim to complete projects with the shortest time, lowest cost, and highest quality. This study proposes a free-parameter multiple objective optimization method to tradeoff time, cost, and quality in non-unit repetitive projects.
INTERNATIONAL JOURNAL OF CONSTRUCTION MANAGEMENT
(2023)
Article
Computer Science, Information Systems
Yu Du, Zhe Wang, Jun Li, Long Shi, Dushantha Nalin K. Jayakody, Quan Chen, Wen Chen, Zhu Han
Summary: In this study, a blockchain-aided distributed edge computing market is proposed, which provides decentralized and verified transaction management. A trustworthiness model is introduced to evaluate the quality of each network entity and a two-level trading mechanism over smart contract is developed for efficient transactions. Additionally, a trustworthiness-driven Proof-of-Stake (PoS) consensus mechanism is put forth to enable verified transactions and fair allocation of block generation reward.
IEEE TRANSACTIONS ON MOBILE COMPUTING
(2023)
Article
Engineering, Electrical & Electronic
Kaitao Meng, Qingqing Wu, Shaodan Ma, Wen Chen, Kunlun Wang, Jun Li
Summary: Driven by UAV's advantages, the existing ISAC system is expected to be revolutionized, but the focus on exploring both functionalities at the same time may ignore practical requirements. Therefore, we propose a new IPSAC mechanism to provide a more flexible trade-off between sensing and communication. By optimizing the UAV trajectory, user association, target sensing selection, and transmit beamforming, while meeting the requirements, we maximize the system achievable rate. Our proposed designs are validated to be effective and unveil a more flexible trade-off over benchmark schemes.
IEEE TRANSACTIONS ON WIRELESS COMMUNICATIONS
(2023)
Article
Computer Science, Information Systems
Yuwen Qian, Wei Li, Yan Lin, Long Shi, Xiangwei Zhou, Jun Li, Feng Shu
Summary: This article proposes a wireless covert communication system where transmit antennas are selected and coded to generate a covert codebook. Covert quadrature amplitude modulation (QAM) and precoding methods are used to modulate and deviate the constellations of covert messages. Optimization techniques are applied to maximize the covert data rate and reduce information leakage.
IEEE INTERNET OF THINGS JOURNAL
(2023)
Article
Engineering, Multidisciplinary
Aoyu Gong, Yijin Zhang, Lei Deng, Fang Liu, Jun Li, Feng Shu
Summary: This paper addresses the dynamic optimization problem of random access in deadline-constrained broadcasting with frame-synchronized traffic. It proposes a dynamic control scheme that allows each active node to determine the transmission probability based on local knowledge. The study develops both a Markov Decision Process (MDP) framework and a Partially Observable MDP (POMDP) framework to obtain optimal or near-optimal schemes in different scenarios. Numerical results validate the proposed approaches.
IEEE TRANSACTIONS ON NETWORK SCIENCE AND ENGINEERING
(2023)
Article
Physics, Multidisciplinary
Cheng Wang, Zhen Mei, Jun Li, Feng Shu, Xuan He, Lingjun Kong
Summary: In this paper, the channel probability density function of 3D NAND flash memory is derived for the first time, considering major sources of errors. The mutual information (MI) is used as a metric to design the quantization, and a dynamic programming algorithm is proposed to jointly optimize read-voltage thresholds for all layers by maximizing the MI (MMI). Furthermore, an MI derivative (MID)-based method is developed to obtain read-voltage thresholds for hard-decision decoding (HDD) of error correction codes (ECCs), reducing complexity.
Article
Telecommunications
Xusheng Zhu, Wen Chen, Zhendong Li, Qingqing Wu, Jun Li
Summary: In this study, a novel quadrature spatial scattering modulation (QSSM) transmission technique based on millimeter wave systems is investigated. The transmitter generates two orthogonal beams to target candidate scatterers in the channel and carry the real and imaginary parts of the conventional signal. The maximum likelihood (ML) detector is used at the receiver to recover the received beams and signals. The closed-form average bit error probability (ABEP) expression and asymptotic ABEP expression of the QSSM scheme are derived and verified through Monte Carlo simulations, demonstrating its improved performance compared to traditional spatial scattering modulation.
IEEE COMMUNICATIONS LETTERS
(2023)
Article
Telecommunications
Shunwai Zhang, Lulu Song, Jun Li, Tho Le-Ngoc
Summary: This paper investigates an energy-efficient decode-and-forward relay cooperation scheme aided by a reconfigurable intelligent surface (RIS) with minimum-rate guarantee. The RIS plays a similar role to traditional relays but has complementary characteristics. Upper bounds on the energy efficiency (EE) are derived for the scheme over Rayleigh fading channels with given transmit powers. The paper addresses the EE optimization problem with minimum-rate guarantee in two scenarios: fixed power and upper-bounded power. Convex optimization techniques are used to solve the equivalent optimal power allocation problem in the fixed power scenario, while fractional programming and generalized Dinkelbach's algorithm are proposed for solving the non-convex EE optimization problem in the upper-bounded power scenario. Simulation results demonstrate the advantages of the considered scheme over benchmark schemes and its robustness against imperfect CSI and discrete phase shifts of RIS.
TELECOMMUNICATION SYSTEMS
(2023)
Article
Computer Science, Information Systems
Zhendong Li, Wen Chen, Ziheng Zhang, Qingqing Wu, Huanqing Cao, Jun Li
Summary: In this article, a state-of-the-art downlink communication transceiver design is proposed for transmissive reconfigurable metasurface (RMS)-enabled simultaneous wireless information and power transfer (SWIPT) networks. The design includes a feed antenna for beamforming and builds spatial propagation models for plane and spherical waves. A robust system sum-rate maximization problem is formulated considering imperfect channel state information (CSI), and the nonconvex optimization problem is solved using an alternating optimization (AO) framework with successive convex approximation (SCA), penalty function method, and difference-of-convex (DC) programming. Numerical results show the proposed algorithm has convergence and outperforms other benchmark algorithms.
IEEE INTERNET OF THINGS JOURNAL
(2023)
Article
Engineering, Electrical & Electronic
Tae-Kyoung Kim, Yo-Seb Jeon, Jun Li, Nima Tavangaran, H. Vincent Poor
Summary: This paper proposes a semi-data-aided channel estimator that improves channel estimation accuracy by utilizing data symbols as pilot signals. The method leverages reinforcement learning to select reliable detected symbols and updates the channel estimate using the selected symbols as additional pilot signals. It defines a Markov decision process (MDP) to determine whether to use each detected symbol and develops a RL algorithm based on Monte Carlo tree search to find an effective policy. The proposed channel estimator requires less computational complexity compared to conventional iterative data-aided channel estimators and effectively mitigates channel estimation error and detection performance loss caused by insufficient pilot signals.
IEEE TRANSACTIONS ON WIRELESS COMMUNICATIONS
(2023)
Proceedings Paper
Computer Science, Artificial Intelligence
Yuan Zhao, Bo Liu, Ming Ding, Baoping Liu, Tianqing Zhu, Xin Yu
Summary: The explosive progress of Deepfake techniques poses unprecedented privacy and security risks to our society by creating real-looking but fake visual content. The current Deepfake detection studies are still in their infancy because they mainly rely on capturing artifacts left by a Deepfake synthesis process as detection clues, which can be easily removed by various distortions (e.g. blurring) or advanced Deepfake techniques.
2023 IEEE/CVF WINTER CONFERENCE ON APPLICATIONS OF COMPUTER VISION (WACV)
(2023)
Proceedings Paper
Computer Science, Artificial Intelligence
Baoping Liu, Bo Liu, Ming Ding, Tianqing Zhu, Xin Yu
Summary: In this paper, the Temporal Identity Inconsistency Network (TI(2)Net) is proposed as a Deepfake detector that focuses on temporal identity inconsistency. It recognizes fake videos by capturing the dissimilarities of human faces among video frames of the same identity. TI(2)Net is a reference-agnostic detector and can be used on unseen datasets.
2023 IEEE/CVF WINTER CONFERENCE ON APPLICATIONS OF COMPUTER VISION (WACV)
(2023)
Proceedings Paper
Computer Science, Software Engineering
Su Yen Chia, Xiwei Xu, Ming Ding, David Smith, Hye-Young Paik, Liming Zhu
Summary: Privacy is a crucial consideration in software systems, and it should be taken into account from the early stages of design to prevent unintended information leakage. Researchers and industry practitioners have explored privacy patterns to address various privacy issues. These patterns serve as reusable design solutions but also pose a challenge in choosing the appropriate ones.
2023 IEEE 20TH INTERNATIONAL CONFERENCE ON SOFTWARE ARCHITECTURE, ICSA
(2023)
Article
Computer Science, Theory & Methods
Xin Yuan, Wei Ni, Ming Ding, Kang Wei, Jun Li, H. Vincent Poor
Summary: This paper proposes a new differential privacy perturbation mechanism with a time-varying noise amplitude to protect the privacy of federated learning and retain the capability of adjusting the learning performance. Extensive experiments confirm the contribution of this mechanism to the convergence and accuracy of privacy-preserving federated learning compared to the state-of-the-art Gaussian noise mechanism with a persistent noise amplitude.
IEEE TRANSACTIONS ON INFORMATION FORENSICS AND SECURITY
(2023)
Article
Engineering, Electrical & Electronic
Ziyan Yin, Zhe Wang, Jun Li, Ming Ding, Wen Chen, Shi Jin
Summary: To address the challenges posed by dynamic and heterogeneous data traffic in 5G and beyond mobile networks, we proposed a learning-based dynamic time-frequency division duplexing (D-TFDD) scheme. By formulating the problem as a decentralized partially observable Markov decision process (Dec-POMDP), we optimized the uplink and downlink time-frequency resource allocation of base stations (BSs) to meet the asymmetric and heterogeneous traffic demands while reducing inter-cell interference. The proposed federated reinforcement learning (RL) algorithm, FWDDPG, enables decentralized global optimization of resource allocation through the exchange of local RL models among neighboring BSs within a federated learning framework. Simulation results demonstrate the superiority of our algorithm in terms of system sum rate compared to benchmark algorithms.
IEEE JOURNAL OF SELECTED TOPICS IN SIGNAL PROCESSING
(2023)