Incentive Mechanism Design for Joint Resource Allocation in Post-Quantum Blockchain-based Federated Learning

Due to the emergence of quantum computers, most of the cryptographic algorithms we currently use can be broken by quantum computers in theory. Therefore, post-quantum cryptographic algorithms that can resist attacks by quantum computers have great promise in blockchain-based federated learning (BCFL). To encourage BCFL participants to use the kyber algorithm (mining and training), we design an incentive mechanism in which the model owner (hereafter referred to as MO) assigns each client an appropriate reward for using the kyber algorithm, a mining reward and a training reward. The clients and MO will use a two-stage Stackelberg game to allocate computational resources for each subtask based on the three rewards given by the MO. We will use lattice cryptography for the full information game.