Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

FBChain: A Blockchain-based Federated Learning Model with Efficiency and Secure Communication

Yang Li, Chunhe Xia, Tianbo Wang

TL;DR

FBChain tackles privacy leakage and inefficient communication in federated learning by storing the global model and local-model hashes on a blockchain, encrypting local parameters, and verifying integrity via hash comparison. It introduces a Proof of Weighted Link Speed (PoWLS) to select high-throughput nodes for global aggregation and block packaging, improving communication efficiency. Experimental results on public datasets show FBChain achieving competitive accuracy with reduced blockchain storage pressure and enhanced transmission performance compared with baseline FL methods. The framework uses a credit score and token reward system to incentivize high-quality contributions, pointing to practical benefits in decentralized FL deployments.

Abstract

Privacy and security in the parameter transmission process of federated learning are currently among the most prominent concerns. However, there are two thorny problems caused by unprotected communication methods: "parameter-leakage" and "inefficient-communication". This article proposes Blockchain-based Federated Learning (FBChain) model for federated learning parameter communication to overcome the above two problems. First, we utilize the immutability of blockchain to store the global model and hash value of local model parameters in case of tampering during the communication process, protect data privacy by encrypting parameters, and verify data consistency by comparing the hash values of local parameters, thus addressing the "parameter-leakage" problem. Second, the Proof of Weighted Link Speed (PoWLS) consensus algorithm comprehensively selects nodes with the higher weighted link speed to aggregate global model and package blocks, thereby solving the "inefficient-communication" problem. Experimental results demonstrate the effectiveness of our proposed FBChain model and its ability to improve model communication efficiency in federated learning.

FBChain: A Blockchain-based Federated Learning Model with Efficiency and Secure Communication

TL;DR

FBChain tackles privacy leakage and inefficient communication in federated learning by storing the global model and local-model hashes on a blockchain, encrypting local parameters, and verifying integrity via hash comparison. It introduces a Proof of Weighted Link Speed (PoWLS) to select high-throughput nodes for global aggregation and block packaging, improving communication efficiency. Experimental results on public datasets show FBChain achieving competitive accuracy with reduced blockchain storage pressure and enhanced transmission performance compared with baseline FL methods. The framework uses a credit score and token reward system to incentivize high-quality contributions, pointing to practical benefits in decentralized FL deployments.

Abstract

Privacy and security in the parameter transmission process of federated learning are currently among the most prominent concerns. However, there are two thorny problems caused by unprotected communication methods: "parameter-leakage" and "inefficient-communication". This article proposes Blockchain-based Federated Learning (FBChain) model for federated learning parameter communication to overcome the above two problems. First, we utilize the immutability of blockchain to store the global model and hash value of local model parameters in case of tampering during the communication process, protect data privacy by encrypting parameters, and verify data consistency by comparing the hash values of local parameters, thus addressing the "parameter-leakage" problem. Second, the Proof of Weighted Link Speed (PoWLS) consensus algorithm comprehensively selects nodes with the higher weighted link speed to aggregate global model and package blocks, thereby solving the "inefficient-communication" problem. Experimental results demonstrate the effectiveness of our proposed FBChain model and its ability to improve model communication efficiency in federated learning.
Paper Structure (15 sections, 4 equations, 5 figures, 1 table)

This paper contains 15 sections, 4 equations, 5 figures, 1 table.

Table of Contents

  1. Introduction
  2. RELATED WORK
  3. Improvement on communication efficiency
  4. Improvement on communication security
  5. BACKGROUND
  6. Symmetric encryption and asymmetric encryption
  7. PoS Consensus algorithm
  8. Blockchain-based Federated Learning Model
  9. Model process
  10. Proof of Weighted Link Speed Consensus Algorithm
  11. PERFORMANCE EVALUATION
  12. Effectiveness of FBChain
  13. Transmission Delay
  14. Credit Score and Stake Trending
  15. Conclusion and Future Work

Figures (5)

  • Figure 1: FBChain Model Architecture
  • Figure 2: Effectiveness of FBChain, VANILLA-FEDAVG, and VBFL in Non-IID Data Distribution
  • Figure 3: Transmission Delay of Local Training Devices, $\kappa = 0, CR_{TH} = 50, CR^r = 5,CR^p = -5$
  • Figure 4: Local Training Devices Credit Score Trends
  • Figure 5: Local Training Devices Stake Trends,$\kappa = 0, CR_{TH} = 50, CR^r = 5,CR^p = -5$