Debunking Optimization Myths in Federated Learning for Medical Image Classification

TL;DR

This paper demonstrates that in federated learning for medical image classification, local device configurations like optimizer and learning rate significantly impact performance, often more than the choice of FL algorithm itself.

Contribution

The study clarifies the influence of edge device configurations on FL performance and benchmarks recent methods, highlighting the importance of proper local setup over algorithm complexity.

Findings

Optimizer and learning rate choices greatly affect FL performance.

Increasing local epochs can improve or impair convergence depending on the method.

Edge-specific configuration is more crucial than algorithm complexity.

Abstract

Federated Learning (FL) is a collaborative learning method that enables decentralized model training while preserving data privacy. Despite its promise in medical imaging, recent FL methods are often sensitive to local factors such as optimizers and learning rates, limiting their robustness in practical deployments. In this work, we revisit vanilla FL to clarify the impact of edge device configurations, benchmarking recent FL methods on colorectal pathology and blood cell classification task. We numerically show that the choice of local optimizer and learning rate has a greater effect on performance than the specific FL method. Moreover, we find that increasing local training epochs can either enhance or impair convergence, depending on the FL method. These findings indicate that appropriate edge-specific configuration is more crucial than algorithmic complexity for achieving effective FL.