Decentralized Federated Learning Privacy-Preserving Framework Based on Blockchain
Abstract
Federated Learning is a promising approach for privacy-preserving distribute training, but it still faces trust issues and privacy concerns. To address such problems, we suggest a decentralized federated learning framework (PP-BFL) based on blockchain that introduces a node reputation value mechanism to establish a trusted reputation system and a training accuracy verification mechanism that leverages smart contracts to ensure model accuracy while protecting data privacy. Our framework reduces the risk of privacy violations due to collusive evil participants and encourages honest and active participation by modeling historical performance and ranking participants based on their trustworthiness. Experimentally, we demonstrate that our framework delivers higher levels of privacy protection. Our work offers an innovative solution for privacy-preserving distribute training, and future work will explore how to optimize and expand our approach to achieve even greater security and privacy guarantees.
Read the paper
Related studies
- Active Learning during Federated Learning for Object Detection
- Fairness-aware federated learning
- Decentralized Federated Learning
- Building a Personally Identifiable Information Recognizer in a Privacy Preserved Manner Using Automated Annotation and Federated Learning
- Preservation of Patient Level Privacy: Federated Classification and Calibration Models
- Efficient explainable federated learning
Go deeper in Evidence Hub
Evidence Hub adds a study data sheet (design, sample size, evidence level, key numbers), an AI summary and a chat grounded in the paper's text.