Proof of Federated Learning: A Novel Energy-Recycling Consensus Algorithm

Proof of work (PoW), the most popular consensus mechanism for blockchain, requires ridiculously large amounts energy but without any useful outcome beyond determining accounting rights among miners. To tackle drawback PoW, we propose a novel energy-recycling algorithm, namely proof federated learning (PoFL), where originally wasted to solve difficult meaningless puzzles in PoW is reinvested learning. Federated and pooled-mining, trend have natural fit terms organization structure. However, separation between data usufruct ownership blockchain lead privacy leakage model training verification, deviating from original intention federal address challenge, reverse game-based trading privacy-preserving verification are proposed. The former can guard against while latter verifies accuracy trained with preservation task requester's test as well pool's submitted model. best our knowledge, article first employ blockchain. Extensive simulations based on synthetic real-world demonstrate effectiveness efficiency proposed mechanisms.