DFPL: Decentralized Federated Prototype Learning Across Heterogeneous Data Distributions

Federated learning is a distributed machine learning paradigm through centralized model aggregation. However, standard federated learning relies on a centralized server, making it vulnerable to server failures. While existing solutions utilize blockchain technology to implement Decentralized Federated Learning (DFL), the statistical heterogeneity of data distributions among clients severely degrades the performance of DFL. Driven by this issue, this paper proposes a decentralized federated prototype learning framework, named DFPL, which significantly improves the performance of DFL across heterogeneous data distributions. Specifically, DFPL introduces prototype learning into DFL to mitigate the impact of statistical heterogeneity and reduces the amount of parameters exchanged between clients. Additionally, blockchain is embedded into our framework, enabling the training and mining processes to be implemented locally on each client. From a theoretical perspective, we analyze the convergence of DFPL by modeling the required computational resources during both training and mining processes. The experiment results highlight the superiority of our DFPL in model performance and communication efficiency across four benchmark datasets with heterogeneous data distributions.