PerFED-GAN: Personalized Federated Learning via Generative Adversarial Networks

TL;DR

PerFED-GAN introduces a personalized federated learning approach using GANs that enables clients to independently design their models without sharing architecture or parameters, significantly improving accuracy amid data and model heterogeneity.

Contribution

It presents a novel federated learning method allowing clients to customize models independently using GANs, addressing model heterogeneity and privacy concerns.

Findings

Outperforms existing methods with 42% higher mean test accuracy.

Effectively handles significant variations in client data and model architectures.

Enhances privacy by avoiding sharing model architecture or parameters.

Abstract

Federated learning is gaining popularity as a distributed machine learning method that can be used to deploy AI-dependent IoT applications while protecting client data privacy and security. Due to the differences of clients, a single global model may not perform well on all clients, so the personalized federated learning method, which trains a personalized model for each client that better suits its individual needs, becomes a research hotspot. Most personalized federated learning research, however, focuses on data heterogeneity while ignoring the need for model architecture heterogeneity. Most existing federated learning methods uniformly set the model architecture of all clients participating in federated learning, which is inconvenient for each client's individual model and local data distribution requirements, and also increases the risk of client model leakage. This paper proposes a federated learning method based on co-training and generative adversarial networks(GANs) that allows each client to design its own model to participate in federated learning training independently without sharing any model architecture or parameter information with other clients or a center. In our experiments, the proposed method outperforms the existing methods in mean test accuracy by 42% when the client's model architecture and data distribution vary significantly.