Fisher-Routed Mixture of Experts for Federated Class-Incremental Learning

2026-06-27

AI Summary

This work addresses key challenges in federated class-incremental learning—namely, model capacity conflicts, catastrophic forgetting, data heterogeneity under non-independent and identically distributed (non-IID) settings, and category misalignment across clients—by introducing a mixture-of-experts framework that achieves an effective balance between knowledge acquisition and retention through adaptive expert specialization and dynamic sample routing. The core contributions include a Fisher information–based expert scoring mechanism (FRES), an adaptive expert selection strategy (AES), and routing-aware regularization (RAR). The proposed method consistently outperforms existing approaches across multiple benchmarks and is theoretically shown to achieve a convergence rate of 𝒪(T⁻¹).

Abstract

Federated Learning (FL) emerged as a promising distributed machine learning paradigm. However, extending FL to the class incremental learning scenarios introduces unique challenges: 1) Capacity conflict and catastrophic forgetting from the shared model overloading, 2) Heterogeneity from Non-Independent and Identically Distributed (Non-IID) data, and 3) Synchronized class misalignment. In this paper, we propose \textbf{F}isher-Routed \textbf{M}i\textbf{X}ture of Experts for \textbf{Fed}erated Class-Incremental Learning (\textsc{FedFMX}), a novel framework to address these challenges via adaptive expert specialization across clients. The crucial insight is to route each sample to an expert subset that jointly optimizes knowledge acquisition and retention. Specifically, we introduce a Fisher-Routed Expert Scoring (FRES) module to estimate expert importance via Fisher-based stability cost and gradient-based plasticity gain. Then, we design an Adaptive Expert Selection (AES) module by quantifying marginal contributions for adaptive expert subset determination. Finally, by the routing-aware regularization (RAR), we achieve load balance and efficient FL training. We theoretically prove the $\mathcal{O}(T^{-1})$ convergence rate. Extensive experiments on multiple benchmarks compared with state-of-the-art methods demonstrate the superiority of \textsc{FedFMX}.

Problem

• Federated Learning
• Class-Incremental Learning
• Catastrophic Forgetting
• Non-IID Data
• Class Misalignment

Innovation

• Fisher-Routed Mixture of Experts
• Federated Class-Incremental Learning
• Expert Routing
• Catastrophic Forgetting
• Non-IID Data