Article
Chemistry, Multidisciplinary
Xiaohui Yang, Zijian Dong
Summary: This paper investigates the issue of data privacy leakage in federated learning and proposes a Kalman Filter-based Differential Privacy Federated Learning Method. Experimental results demonstrate that the proposed method outperforms traditional differential privacy federated learning in terms of accuracy.
APPLIED SCIENCES-BASEL
(2022)
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
Zunming Chen, Hongyan Cui, Ensen Wu, Xi Yu
Summary: With the proliferation of the Internet of Things (IoT) and the use of devices with sensing, computing, and communication capabilities, intelligent applications empowered by artificial intelligence have become prevalent. However, existing classical artificial intelligence algorithms face challenges in realistic intelligent IoT applications due to data privacy concerns and distributed datasets. To address this, the paper proposes a novel efficient adaptive federated optimization (FedEAFO) algorithm that minimizes learning error by jointly considering local update and parameter compression variables to improve the efficiency of Federated Learning (FL). Experimental results demonstrate that FedEAFO achieves higher accuracies faster compared to state-of-the-art algorithms.
Article
Physics, Multidisciplinary
Xiaoying Shen, Hang Jiang, Yange Chen, Baocang Wang, Le Gao
Summary: This paper introduces a perturbation algorithm (PDPM) that satisfies personalized local differential privacy (PLDP), resolving the issue of inadequate or excessive privacy protection for some participants due to the same privacy budget set for all clients.
Article
Computer Science, Hardware & Architecture
Wenyan Liu, Junhong Cheng, Xiaoling Wang, Xingjian Lu, Jianwei Yin
Summary: This paper proposes a secure and reliable federated learning algorithm by integrating hybrid differential privacy into federated learning. The algorithm divides users into two categories according to their different privacy needs, and introduces an adaptive gradient clip scheme and an improved composition method to reduce the effects of noise and clip.
JOURNAL OF SYSTEMS ARCHITECTURE
(2022)
Article
Computer Science, Hardware & Architecture
Xiaolan Tang, Yuting Liang, Guan Wang, Wenlong Chen
Summary: This article introduces an assisted driving system based on federated reinforcement learning, which can help visually impaired persons safely control vehicles while protecting their privacy data.
Article
Computer Science, Information Systems
Baocang Wang, Yange Chen, Hang Jiang, Zhen Zhao
Summary: This article proposes a novel privacy-preserving edge FL framework based on LDP (PPeFL), which addresses the privacy issues in FL through three LDP mechanisms. Extensive experiments demonstrate the practicality and efficiency of the PPeFL scheme.
IEEE INTERNET OF THINGS JOURNAL
(2023)
Article
Computer Science, Artificial Intelligence
Chen Zhang, Yu Xie, Hang Bai, Bin Yu, Weihong Li, Yuan Gao
Summary: Federated learning is a setup where multiple clients collaborate to solve machine learning problems under the coordination of a central aggregator. It reduces systematic privacy risks and costs through local computing and model transmission. This method ensures data privacy for each device and improves learning efficiency and security.
KNOWLEDGE-BASED SYSTEMS
(2021)
Article
Computer Science, Information Systems
Ahmed El Ouadrhiri, Ahmed Abdelhadi
Summary: This article introduces the main ideas of differential privacy in protecting user privacy during the deep learning process, including the use of noise to protect the original data and learning parameters. The study also discusses various types of probability distributions that satisfy the differential privacy mechanism and provides an overview of different variants of differential privacy.
Article
Computer Science, Information Systems
M. A. P. Chamikara, P. Bertok, I. Khalil, D. Liu, S. Camtepe
Summary: Edge computing and distributed machine learning have advanced to a level that can revolutionize a particular organization. Federated learning (FedML) is a method to protect privacy in machine learning, but additional measures are needed to ensure data privacy.
COMPUTER COMMUNICATIONS
(2021)
Article
Computer Science, Information Systems
Bowen Zhao, Ximeng Liu, Wei-Neng Chen, Robert H. Deng
Summary: In this paper, a privacy-preserving mobile crowdsensing system called CROWDFL is proposed by integrating federated learning (FL) into MCS. Participants in CROWDFL locally process sensing data and only upload encrypted training models to the server to protect their privacy. The system also includes a secure aggregation algorithm and a hybrid incentive mechanism to improve efficiency and stimulate participation.
IEEE TRANSACTIONS ON MOBILE COMPUTING
(2023)
Article
Automation & Control Systems
Shuzhen Chen, Dongxiao Yu, Yifei Zou, Jiguo Yu, Xiuzhen Cheng
Summary: This article studies decentralized federated learning algorithm DWFL in wireless IoT networks, which organizes workers in a peer-to-peer and server-less manner, exchanging privacy preserving data with analog transmission scheme over wireless channels in parallel, achieving good privacy protection and convergence rate.
IEEE TRANSACTIONS ON INDUSTRIAL INFORMATICS
(2022)
Article
Automation & Control Systems
Zehui Zhang, Linlin Zhang, Qingdan Li, Kunshu Wang, Ningxin He, Tiegang Gao
Summary: This paper proposes a Privacy-Enhanced Momentum Federated Learning framework (PEMFL) that protects the privacy information of industrial agents through the use of differential privacy and chaos-based encryption method. Experimental results demonstrate that PEMFL performs well in terms of accuracy and privacy security.
Article
Computer Science, Artificial Intelligence
Laraib Javed, Adeel Anjum, Bello Musa Yakubu, Majid Iqbal, Syed Atif Moqurrab, Gautam Srivastava
Summary: In our technologically advanced world, data security is crucial for every individual, especially in the exchange of medical information. Existing techniques for preserving data security have limitations, but this work proposes a secure architecture and semantic approach based on blockchain, local differential privacy, and federated learning. The proposed framework addresses the vulnerabilities of current solutions and provides a trustless environment for data sharing.
Article
Computer Science, Information Systems
Ming Yang, Hang Cheng, Fei Chen, Ximeng Liu, Meiqing Wang, Xibin Li
Summary: Although federated learning provides privacy protection, studies have shown that shared parameters or gradients may still reveal user privacy. Differential privacy offers a promising solution with low computational overhead. However, it also introduces the risk of model poisoning attacks, where attackers can manipulate the model using noise to reduce convergence speed and cause divergence.
INFORMATION SCIENCES
(2023)