| 注册
首页|期刊导航|计算机工程与应用|基于区块链的联邦学习系统方案研究综述

基于区块链的联邦学习系统方案研究综述

郭思昀 李雷孝 杜金泽 林浩

计算机工程与应用2025,Vol.61Issue(15):36-53,18.
计算机工程与应用2025,Vol.61Issue(15):36-53,18.DOI:10.3778/j.issn.1002-8331.2410-0227

基于区块链的联邦学习系统方案研究综述

Survey of Blockchain-Based Federated Learning System Schemes

郭思昀 1李雷孝 2杜金泽 3林浩3

作者信息

  • 1. 内蒙古工业大学 数据科学与应用学院,呼和浩特 010080
  • 2. 内蒙古工业大学 数据科学与应用学院,呼和浩特 010080||内蒙古自治区北疆网络空间安全重点实验室,呼和浩特 010080||内蒙古自治区基于大数据的软件服务工程技术研究中心,呼和浩特 010080
  • 3. 天津理工大学 计算机科学与工程学院,天津 300384
  • 折叠

摘要

Abstract

Federated learning enables multiple participants to collaboratively train models while protecting data privacy,and it has been widely applied in various fields such as healthcare and smart cities,making significant contributions to breaking the"data silo"phenomenon.However,in practical applications,federated learning faces issues such as single points of failure,susceptibility to malicious attacks,and data privacy breaches.To address these,researchers have consid-ered leveraging blockchain technology,which is characterized by decentralization,immutability,and high transparency,to provide a secure data exchange platform for federated learning.However,the integrated architecture still encounters chal-lenges such as the lack of effective incentive mechanisms,high computational and storage costs,and the absence of mali-cious model detection mechanisms.Through in-depth research and exploration of the potential and challenges of block-chain-based federated learning systems,this paper discusses the optimization and improvement of consensus mechanisms,privacy protection schemes,network security measures,incentive mechanisms,and secure aggregation methods within the architecture,starting from the existing problems in different frameworks and technologies.Subsequently,the paper elaborates on the key role of smart contracts in automated execution,contribution assessment,and reward distribution.Finally,it identifies challenges in client selection,data heterogeneity processing,privacy protection,reward distribution,and blockchain pain points,and provides an outlook on future research directions.

关键词

去中心化联邦学习/区块链/隐私保护/激励机制/智能合约

Key words

decentralized federated learning/blockchain/privacy protection/incentive mechanism/smart contract

分类

信息技术与安全科学

引用本文复制引用

郭思昀,李雷孝,杜金泽,林浩..基于区块链的联邦学习系统方案研究综述[J].计算机工程与应用,2025,61(15):36-53,18.

基金项目

国家自然科学基金(62362055) (62362055)

内蒙古自治区重点研发与成果转化计划项目(2022YFSJ0013,2023YFHH0052) (2022YFSJ0013,2023YFHH0052)

内蒙古自治区高等学校青年科技英才支持计划项目(NJYT22084,NJYT24035) (NJYT22084,NJYT24035)

内蒙古自然科学基金(2023MS06008) (2023MS06008)

内蒙古自治区直属高校科研项目(JY20220061,JY20230119,JY20230019) (JY20220061,JY20230119,JY20230019)

鄂尔多斯市重点研发计划项目(YF20232328) (YF20232328)

国家重点研发项目(2023YFB2703900). (2023YFB2703900)

计算机工程与应用

OA北大核心

1002-8331

访问量0
|
下载量0
段落导航相关论文