Generative Data Augmentation for Non-IID Problem in Decentralized Clinical Machine Learning

TL;DR

This paper introduces SL-GAN, a generative data augmentation framework for swarm learning that improves performance on non-IID clinical data by generating synthetic data, with proven convergence and superior results on real datasets.

Contribution

SL-GAN is the first to integrate generative augmentation into decentralized swarm learning for non-IID data, with theoretical convergence guarantees.

Findings

SL-GAN outperforms existing methods on clinical datasets.

Theoretical proof of convergence for SL-GAN.

Effective augmentation improves model performance in non-IID settings.

Abstract

Swarm learning (SL) is an emerging promising decentralized machine learning paradigm and has achieved high performance in clinical applications. SL solves the problem of a central structure in federated learning by combining edge computing and blockchain-based peer-to-peer network. While there are promising results in the assumption of the independent and identically distributed (IID) data across participants, SL suffers from performance degradation as the degree of the non-IID data increases. To address this problem, we propose a generative augmentation framework in swarm learning called SL-GAN, which augments the non-IID data by generating the synthetic data from participants. SL-GAN trains generators and discriminators locally, and periodically aggregation via a randomly elected coordinator in SL network. Under the standard assumptions, we theoretically prove the convergence of SL-GAN using stochastic approximations. Experimental results demonstrate that SL-GAN outperforms state-of-art methods on three real world clinical datasets including Tuberculosis, Leukemia, COVID-19.