FedCCL: Federated Dual-Clustered Feature Contrast Under Domain Heterogeneity

TL;DR

This paper introduces FedCCL, a federated learning framework that uses dual clustering of local and global features to better handle non-IID data heterogeneity, improving model performance.

Contribution

It proposes a novel dual-clustered feature contrast approach that captures intra- and inter-client semantic similarities for federated learning with non-IID data.

Findings

Achieves superior performance on multiple datasets.

Effectively handles intra-domain and inter-domain heterogeneity.

Improves model robustness in federated settings.

Abstract

Federated learning (FL) facilitates a privacy-preserving neural network training paradigm through collaboration between edge clients and a central server. One significant challenge is that the distributed data is not independently and identically distributed (non-IID), typically including both intra-domain and inter-domain heterogeneity. However, recent research is limited to simply using averaged signals as a form of regularization and only focusing on one aspect of these non-IID challenges. Given these limitations, this paper clarifies these two non-IID challenges and attempts to introduce cluster representation to address them from both local and global perspectives. Specifically, we propose a dual-clustered feature contrast-based FL framework with dual focuses. First, we employ clustering on the local representations of each client, aiming to capture intra-class information based on these local clusters at a high level of granularity. Then, we facilitate cross-client knowledge sharing by pulling the local representation closer to clusters shared by clients with similar semantics while pushing them away from clusters with dissimilar semantics. Second, since the sizes of local clusters belonging to the same class may differ for each client, we further utilize clustering on the global side and conduct averaging to create a consistent global signal for guiding each local training in a contrastive manner. Experimental results on multiple datasets demonstrate that our proposal achieves comparable or superior performance gain under intra-domain and inter-domain heterogeneity.