Proof-of-Collaborative-Learning: A Multi-winner Federated Learning Consensus Algorithm