Latency Optimization for Blockchain-Empowered Federated Learning in Multi-Server Edge Computing

In this paper, we study a new latency optimization problem for blockchain-based federated learning (BFL) in multi-server edge computing. system model, distributed mobile devices (MDs) communicate with set of servers (ESs) to handle both machine (ML) model training and block mining simultaneously. To assist the ML resource-constrained MDs, develop an offloading strategy that enables MDs transmit their data one associated ESs. We then propose decentralized aggregation solution at layer based on consensus mechanism build global via peer-to-peer (P2P)-based blockchain communications. Blockchain builds trust among ESs facilitate reliable sharing cooperative formation, rapid elimination manipulated models caused by poisoning attacks. formulate latency-aware BFL as aiming minimize joint consideration decisions, MDs’ power, channel bandwidth allocation offloading, computational allocation, hash power allocation. Given mixed action space discrete continuous variables, novel deep reinforcement scheme parameterized advantage actor critic algorithm. theoretically characterize convergence properties terms delay, mini-batch size, number P2P communication rounds. Our numerical evaluation demonstrates superiority our proposed over baselines efficiency, rate, latency, robustness against