FMCL: Class-Aware Client Clustering with Foundation Model Representations for Heterogeneous Federated Learning

arXiv cs.LG / 5/1/2026

📰 NewsIdeas & Deep AnalysisModels & Research

共有:

Key Points

The paper introduces FMCL, a one-shot, class-aware client clustering method for federated learning under statistical heterogeneity.
FMCL uses a frozen foundation model to compute class-level embedding prototypes for each client and clusters clients by cosine similarity of their class-aware representations.
By performing clustering only once before training, FMCL avoids iterative coordination and adds no extra communication during federated optimization.
Experiments on heterogeneous benchmarks show that FMCL improves federated performance and produces more stable client clustering than prior clustering approaches under non-IID data.
The approach is designed to be architecture-agnostic for downstream federated models, improving its practical adaptability.

Abstract

Federated Learning (FL) enables collaborative model training across distributed clients without sharing raw data, yet its performance deteriorates under statistical heterogeneity. Clustered Federated Learning addresses this challenge by grouping similar clients and training separate models per cluster. However, existing clustering strategies often rely on raw data statistics, model parameters, or heuristic similarity measures that fail to capture class-level semantic structure across heterogeneous domains and frequently require iterative coordination. We propose FMCL, a one-shot, class-aware client clustering framework that leverages foundation model representations to construct semantic client signatures. Using a frozen foundation model, FMCL computes class-level embedding prototypes for each client and measures similarity via cosine distance between their class-aware representations. Clustering is performed once prior to training, introducing no additional communication during federated optimization and remaining agnostic to the downstream model architecture. Extensive experiments across heterogeneous benchmarks demonstrate that FMCL improves federated performance and yields more stable clustering behavior compared to existing clustering-based methods under non-identically distributed data partitioning.