Privacy-preserving Decentralized Federated Learning over Time-varying Communication Graph
This addresses the challenge of secure and adaptive decentralized learning for mobile or dynamic networks, though it appears incremental as it combines existing techniques like consensus methods and secret sharing.
The paper tackles the problem of enabling privacy-preserving federated learning in a fully decentralized, peer-to-peer setting with time-varying communication graphs, and proposes an algorithm that integrates Metropolis-Hastings and Shamir's secret sharing to achieve this, with correctness and privacy properties established and computational efficiency evaluated via simulation on a real-world dataset.
Establishing how a set of learners can provide privacy-preserving federated learning in a fully decentralized (peer-to-peer, no coordinator) manner is an open problem. We propose the first privacy-preserving consensus-based algorithm for the distributed learners to achieve decentralized global model aggregation in an environment of high mobility, where the communication graph between the learners may vary between successive rounds of model aggregation. In particular, in each round of global model aggregation, the Metropolis-Hastings method is applied to update the weighted adjacency matrix based on the current communication topology. In addition, the Shamir's secret sharing scheme is integrated to facilitate privacy in reaching consensus of the global model. The paper establishes the correctness and privacy properties of the proposed algorithm. The computational efficiency is evaluated by a simulation built on a federated learning framework with a real-word dataset.