Federated Learning with Extremely Noisy Clients via Negative Distillation

• URL: http://arxiv.org/abs/2312.12703v2
• Date: Fri, 23 Feb 2024 02:50:18 GMT
• Title: Federated Learning with Extremely Noisy Clients via Negative Distillation
• Authors: Yang Lu, Lin Chen, Yonggang Zhang, Yiliang Zhang, Bo Han, Yiu-ming Cheung, Hanzi Wang
• Abstract summary: Federated learning (FL) has shown remarkable success in cooperatively training deep models, while struggling with noisy labels. We propose a novel approach, called negative distillation (FedNed) to leverage models trained on noisy clients. FedNed first identifies noisy clients and employs rather than discards the noisy clients in a knowledge distillation manner.
• Score: 70.13920804879312
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Federated learning (FL) has shown remarkable success in cooperatively training deep models, while typically struggling with noisy labels. Advanced works propose to tackle label noise by a re-weighting strategy with a strong assumption, i.e., mild label noise. However, it may be violated in many real-world FL scenarios because of highly contaminated clients, resulting in extreme noise ratios, e.g., $>$90%. To tackle extremely noisy clients, we study the robustness of the re-weighting strategy, showing a pessimistic conclusion: minimizing the weight of clients trained over noisy data outperforms re-weighting strategies. To leverage models trained on noisy clients, we propose a novel approach, called negative distillation (FedNed). FedNed first identifies noisy clients and employs rather than discards the noisy clients in a knowledge distillation manner. In particular, clients identified as noisy ones are required to train models using noisy labels and pseudo-labels obtained by global models. The model trained on noisy labels serves as a `bad teacher' in knowledge distillation, aiming to decrease the risk of providing incorrect information. Meanwhile, the model trained on pseudo-labels is involved in model aggregation if not identified as a noisy client. Consequently, through pseudo-labeling, FedNed gradually increases the trustworthiness of models trained on noisy clients, while leveraging all clients for model aggregation through negative distillation. To verify the efficacy of FedNed, we conduct extensive experiments under various settings, demonstrating that FedNed can consistently outperform baselines and achieve state-of-the-art performance. Our code is available at https://github.com/linChen99/FedNed.

Related papers

• Federated Learning Client Pruning for Noisy Labels [6.30126491637621]
  Federated Learning (FL) enables collaborative model training across decentralized edge devices. This paper introduces ClipFL, a novel framework addressing noisy labels from a fresh perspective. It identifies and excludes noisy clients based on their performance on a clean validation dataset.
  arXiv  Detail & Related papers  (2024-11-11T21:46:34Z)
• FedDiv: Collaborative Noise Filtering for Federated Learning with Noisy Labels [99.70895640578816]
  Federated learning with noisy labels (F-LNL) aims at seeking an optimal server model via collaborative distributed learning. We present FedDiv to tackle the challenges of F-LNL. Specifically, we propose a global noise filter called Federated Noise Filter.
  arXiv  Detail & Related papers  (2023-12-19T15:46:47Z)
• Label Noise-Robust Learning using a Confidence-Based Sieving Strategy [15.997774467236352]
  In learning tasks with label noise, improving model robustness against overfitting is a pivotal challenge. Identifying the samples with noisy labels and preventing the model from learning them is a promising approach to address this challenge. We propose a novel discriminator metric called confidence error and a sieving strategy called CONFES to differentiate between the clean and noisy samples effectively.
  arXiv  Detail & Related papers  (2022-10-11T10:47:28Z)
• Neighborhood Collective Estimation for Noisy Label Identification and Correction [92.20697827784426]
  Learning with noisy labels (LNL) aims at designing strategies to improve model performance and generalization by mitigating the effects of model overfitting to noisy labels. Recent advances employ the predicted label distributions of individual samples to perform noise verification and noisy label correction, easily giving rise to confirmation bias. We propose Neighborhood Collective Estimation, in which the predictive reliability of a candidate sample is re-estimated by contrasting it against its feature-space nearest neighbors.
  arXiv  Detail & Related papers  (2022-08-05T14:47:22Z)
• FedCorr: Multi-Stage Federated Learning for Label Noise Correction [80.9366438220228]
  Federated learning (FL) is a privacy-preserving distributed learning paradigm that enables clients to jointly train a global model. We propose $textttFedCorr$, a general multi-stage framework to tackle heterogeneous label noise in FL. Experiments conducted on CIFAR-10/100 with federated synthetic label noise, and on a real-world noisy dataset, Clothing1M, demonstrate that $textttFedCorr$ is robust to label noise.
  arXiv  Detail & Related papers  (2022-04-10T12:51:18Z)
• Open-set Label Noise Can Improve Robustness Against Inherent Label Noise [27.885927200376386]
  We show that open-set noisy labels can be non-toxic and even benefit the robustness against inherent noisy labels. We propose a simple yet effective regularization by introducing Open-set samples with Dynamic Noisy Labels (ODNL) into training.
  arXiv  Detail & Related papers  (2021-06-21T07:15:50Z)
• Tackling Instance-Dependent Label Noise via a Universal Probabilistic Model [80.91927573604438]
  This paper proposes a simple yet universal probabilistic model, which explicitly relates noisy labels to their instances. Experiments on datasets with both synthetic and real-world label noise verify that the proposed method yields significant improvements on robustness.
  arXiv  Detail & Related papers  (2021-01-14T05:43:51Z)
• Deep k-NN for Noisy Labels [55.97221021252733]
  We show that a simple $k$-nearest neighbor-based filtering approach on the logit layer of a preliminary model can remove mislabeled data and produce more accurate models than many recently proposed methods.
  arXiv  Detail & Related papers  (2020-04-26T05:15:36Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.