FedFed: Feature Distillation against Data Heterogeneity in Federated Learning

TL;DR

FedFed introduces a feature distillation method in federated learning that shares only performance-sensitive features across clients to address data heterogeneity while preserving privacy.

Contribution

This work proposes FedFed, a novel federated learning approach that partitions features into sensitive and robust types, sharing only the sensitive features to improve performance amidst data heterogeneity.

Findings

FedFed improves model performance in heterogeneous data settings.

Sharing only performance-sensitive features reduces privacy risks.

Experiments validate the effectiveness of FedFed across datasets.

Abstract

Federated learning (FL) typically faces data heterogeneity, i.e., distribution shifting among clients. Sharing clients' information has shown great potentiality in mitigating data heterogeneity, yet incurs a dilemma in preserving privacy and promoting model performance. To alleviate the dilemma, we raise a fundamental question: \textit{Is it possible to share partial features in the data to tackle data heterogeneity?} In this work, we give an affirmative answer to this question by proposing a novel approach called {\textbf{Fed}erated \textbf{Fe}ature \textbf{d}istillation} (FedFed). Specifically, FedFed partitions data into performance-sensitive features (i.e., greatly contributing to model performance) and performance-robust features (i.e., limitedly contributing to model performance). The performance-sensitive features are globally shared to mitigate data heterogeneity, while the performance-robust features are kept locally. FedFed enables clients to train models over local and shared data. Comprehensive experiments demonstrate the efficacy of FedFed in promoting model performance.