mlsanityData to 7 Oct 2026

Learning Unlabeled Clients Divergence for Federated Semi-Supervised Learning via Anchor Model Aggregation

arXiv 2407.10327 ↗First posted 14 Jul 2024cs.LG cs.AI cs.CVcited by 0

Not on the map. Why?

Abstract (arXiv)

Federated semi-supervised learning (FedSemi) refers to scenarios where there may be clients with fully labeled data, clients with partially labeled, and even fully unlabeled clients while preserving data privacy. However, challenges arise from client drift due to undefined heterogeneous class distributions and erroneous pseudo-labels. Existing FedSemi methods typically fail to aggregate models from unlabeled clients due to their inherent unreliability, thus overlooking unique information from their heterogeneous data distribution, leading to sub-optimal results. In this paper, we enable unlabeled client aggregation through SemiAnAgg, a novel Semi-supervised Anchor-Based federated Aggregation. SemiAnAgg learns unlabeled client contributions via an anchor model, effectively harnessing their informative value. Our key idea is that by feeding local client data to the same global model and the same consistently initialized anchor model (i.e., random model), we can measure the importance of each unlabeled client accordingly. Extensive experiments demonstrate that SemiAnAgg achieves new state-of-the-art results on four widely used FedSemi benchmarks, leading to substantial performance improvements: a 9% increase in accuracy on CIFAR-100 and a 7.6% improvement in recall on the medical dataset ISIC-18, compared with prior state-of-the-art. Code is available at: https://github.com/xmed-lab/SemiAnAgg.

Key facts. No citing papers in the map yet (data through 7 Oct 2026). Builds on 44 of its 58 references that are in the map. First posted on arXiv 14 Jul 2024.

Builds on

44 of 58 references.

References · 44

  1. 2015Deep Residual Learning for Image Recognitioncited by 51,809
  2. 2019PyTorch: An Imperative Style, High-Performance Deep Learning Librarycited by 20,160
  3. 2016Communication-Efficient Learning of Deep Networks from Decentralized Datacited by 7,592
  4. 2015Delving Deep into Rectifiers: Surpassing Human-Level Performance on ImageNet Classificationcited by 3,590
  5. 2019RandAugment: Practical automated data augmentation with a reduced search spacecited by 2,261
  6. 2020FixMatch: Simplifying Semi-Supervised Learning with Consistency and Confidencecited by 2,080
  7. 2021Barlow Twins: Self-Supervised Learning via Redundancy Reductioncited by 1,818
  8. 2016Temporal Ensembling for Semi-Supervised Learningcited by 1,054
  9. 2021Model-Contrastive Federated Learningcited by 762
  10. 2019Data Shapley: Equitable Valuation of Data for Machine Learningcited by 617
  11. 2020Federated Learning with Matched Averagingcited by 564
  12. 2021FlexMatch: Boosting Semi-Supervised Learning with Curriculum Pseudo Labelingcited by 537
32 more
  1. 2020Reliable Fidelity and Diversity Metrics for Generative Modelscited by 425
  2. 2020Balanced Meta-Softmax for Long-Tailed Visual Recognitioncited by 397
  3. 2015Semi-Supervised Learning with Ladder Networkscited by 246
  4. 2022FreeMatch: Self-adaptive Thresholding for Semi-supervised Learningcited by 232
  5. 2020Rethinking the Value of Labels for Improving Class-Imbalanced Learningcited by 216
  6. 2021CReST: A Class-Rebalancing Self-Training Framework for Imbalanced Semi-Supervised Learningcited by 174
  7. 2021Delving into Deep Imbalanced Regressioncited by 157
  8. 2019Transformation Consistent Self-ensembling Model for Semi-supervised Medical Image Segmentationcited by 138
  9. 2023SoftMatch: Addressing the Quantity-Quality Trade-off in Semi-supervised Learningcited by 133
  10. 2020Federated Semi-Supervised Learning with Inter-Client Consistency & Disjoint Learningcited by 112
  11. 2020Federated Unsupervised Representation Learningcited by 92
  12. 2022Divergence-aware Federated Self-Supervised Learningcited by 71
  13. 2023FedDisco: Federated Learning with Discrepancy-Aware Collaborationcited by 66
  14. 2023FedFed: Feature Distillation against Data Heterogeneity in Federated Learningcited by 54
  15. 2021DASO: Distribution-Aware Semantics-Oriented Pseudo-label for Imbalanced Semi-Supervised Learningcited by 51
  16. 2023Fair Federated Medical Image Segmentation via Client Contribution Estimationcited by 50
  17. 2021SemiFL: Semi-Supervised Federated Learning for Unlabeled Clients with Alternate Trainingcited by 45
  18. 2022Orchestra: Unsupervised Federated Learning via Globally Consistent Clusteringcited by 45
  19. 2021Federated Semi-supervised Medical Image Classification via Inter-client Relation Matchingcited by 42
  20. 2022Label-Efficient Self-Supervised Federated Learning for Tackling Data Heterogeneity in Medical Imagingcited by 39
  21. 2023Towards Generic Semi-Supervised Framework for Volumetric Medical Image Segmentationcited by 35
  22. 2022Invariant Feature Learning for Generalized Long-Tailed Classificationcited by 34
  23. 2022RSCFed: Random Sampling Consensus Federated Semi-supervised Learningcited by 32
  24. 2021SemiFed: Semi-supervised Federated Learning with Consistency and Pseudo-Labelingcited by 29
  25. 2023DHC: Dual-debiased Heterogeneous Co-training Framework for Class-imbalanced Semi-supervised Medical Image Segmentationcited by 25
  26. 2023Semi-supervised learning made simple with self-supervised clusteringcited by 20
  27. 2023ProtoFL: Unsupervised Federated Learning via Prototypical Distillationcited by 11
  28. 2022An Embarrassingly Simple Approach to Semi-Supervised Few-Shot Learningcited by 10
  29. 2023Towards Unbiased Training in Federated Open-world Semi-supervised Learningcited by 6
  30. 2023Local or Global: Selective Knowledge Assimilation for Federated Learning with Limited Labelscited by 4
  31. 2023Federated Model Aggregation via Self-Supervised Priors for Highly Imbalanced Medical Image Classificationcited by 4
  32. 2023Rethinking Semi-Supervised Federated Learning: How to co-train fully-labeled and fully-unlabeled client imaging datacited by 4

Built on by

Not cited yet.