FedSIR: Spectral Client Identification and Relabeling for Federated Learning with Noisy Labels
Fuente:
arXiv
Saved in:
| Main Authors: | , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2026
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866917429883437056 |
|---|---|
| author | Gholami, Sina Ali, Abdulmoneam Haghighi, Tania Arafa, Ahmed Alam, Minhaj Nur |
| author_facet | Gholami, Sina Ali, Abdulmoneam Haghighi, Tania Arafa, Ahmed Alam, Minhaj Nur |
| contents | 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. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2604_20825 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | FedSIR: Spectral Client Identification and Relabeling for Federated Learning with Noisy Labels Gholami, Sina Ali, Abdulmoneam Haghighi, Tania Arafa, Ahmed Alam, Minhaj Nur Machine Learning Artificial Intelligence Computer Vision and Pattern Recognition Distributed, Parallel, and Cluster Computing Signal Processing 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. |
| title | FedSIR: Spectral Client Identification and Relabeling for Federated Learning with Noisy Labels |
| topic | Machine Learning Artificial Intelligence Computer Vision and Pattern Recognition Distributed, Parallel, and Cluster Computing Signal Processing |
| url | https://arxiv.org/abs/2604.20825 |