FedSIR: Spectral Client Identification and Relabeling for Federated Learning with Noisy Labels

2026-04-22

AI Summary

This work addresses the significant degradation in model performance caused by label noise on client devices in federated learning. To tackle this challenge, the authors propose FedSIR, a novel framework that introduces spectral consistency analysis into federated learning for the first time. By characterizing the spectral structure of client feature subspaces, FedSIR enables efficient client-level noise identification and sample-level relabeling with low communication overhead. The approach integrates a spectral reference from clean clients, logit-adjusted loss, knowledge distillation, and distance-aware aggregation to form a comprehensive noise-aware training strategy. Evaluated on standard federated learning benchmarks, FedSIR consistently outperforms existing methods, demonstrating substantial improvements in model accuracy and stability under label noise conditions.

Technology Category

• Machine Learning: Distributed Machine Learning & Federated Learning
• Computer Vision: Adversarial Attacks & Robustness
• Natural Language Processing: Safety and Robustness

Application Category

• User Modeling, Personalization and Recommendation: Federated recommendation systems and personalization
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs

Abstract

Federated learning (FL) enables collaborative model training without sharing raw data; however, the presence of noisy labels across distributed clients can severely degrade the learning performance. In this paper, we propose FedSIR, a multi-stage framework for robust FL under noisy labels. Different from existing approaches that mainly rely on designing noise-tolerant loss functions or exploiting loss dynamics during training, our method leverages the spectral structure of client feature representations to identify and mitigate label noise. Our framework consists of three key components. First, we identify clean and noisy clients by analyzing the spectral consistency of class-wise feature subspaces with minimal communication overhead. Second, clean clients provide spectral references that enable noisy clients to relabel potentially corrupted samples using both dominant class directions and residual subspaces. Third, we employ a noise-aware training strategy that integrates logit-adjusted loss, knowledge distillation, and distance-aware aggregation to further stabilize federated optimization. Extensive experiments on standard FL benchmarks demonstrate that FedSIR consistently outperforms state-of-the-art methods for FL with noisy labels. The code is available at https://github.com/sinagh72/FedSIR.

Problem

• Federated Learning
• Noisy Labels
• Client Identification
• Label Relabeling
• Spectral Analysis

Innovation

• spectral analysis
• federated learning
• label noise
• client relabeling
• feature subspace