mlsanityData to 7 Oct 2026

FlexMatch: Boosting Semi-Supervised Learning with Curriculum Pseudo Labeling

arXiv 2110.08263 ↗First posted 15 Oct 2021cs.LG cs.CVcited by 537

Not on the map. Why?

Abstract (arXiv)

The recently proposed FixMatch achieved state-of-the-art results on most semi-supervised learning (SSL) benchmarks. However, like other modern SSL algorithms, FixMatch uses a pre-defined constant threshold for all classes to select unlabeled data that contribute to the training, thus failing to consider different learning status and learning difficulties of different classes. To address this issue, we propose Curriculum Pseudo Labeling (CPL), a curriculum learning approach to leverage unlabeled data according to the model's learning status. The core of CPL is to flexibly adjust thresholds for different classes at each time step to let pass informative unlabeled data and their pseudo labels. CPL does not introduce additional parameters or computations (forward or backward propagation). We apply CPL to FixMatch and call our improved algorithm FlexMatch. FlexMatch achieves state-of-the-art performance on a variety of SSL benchmarks, with especially strong performances when the labeled data are extremely limited or when the task is challenging. For example, FlexMatch achieves 13.96% and 18.96% error rate reduction over FixMatch on CIFAR-100 and STL-10 datasets respectively, when there are only 4 labels per class. CPL also significantly boosts the convergence speed, e.g., FlexMatch can use only 1/5 training time of FixMatch to achieve even better performance. Furthermore, we show that CPL can be easily adapted to other SSL algorithms and remarkably improve their performances. We open-source our code at https://github.com/TorchSSL/TorchSSL.

Key facts. Cited by 537 papers in the map (data through 7 Oct 2026). Builds on 24 of its 47 references that are in the map. First posted on arXiv 15 Oct 2021.

Builds on

24 of 47 references.

References · 24

  1. 2015Deep Residual Learning for Image Recognitioncited by 51,809
  2. 2019PyTorch: An Imperative Style, High-Performance Deep Learning Librarycited by 20,160
  3. 2016SGDR: Stochastic Gradient Descent with Warm Restartscited by 4,993
  4. 2016Wide Residual Networkscited by 3,057
  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. 2019MixMatch: A Holistic Approach to Semi-Supervised Learningcited by 1,390
  8. 2019Self-training with Noisy Student improves ImageNet classificationcited by 1,304
  9. 2019Unsupervised Data Augmentation for Consistency Trainingcited by 1,057
  10. 2016Temporal Ensembling for Semi-Supervised Learningcited by 1,054
  11. 2017Virtual Adversarial Training: A Regularization Method for Supervised and Semi-Supervised Learningcited by 918
  12. 2019Pseudo-Labeling and Confirmation Bias in Deep Semi-Supervised Learningcited by 507
12 more
  1. 2016Regularization With Stochastic Transformations and Perturbations for Deep Semi-Supervised Learningcited by 441
  2. 2021In Defense of Pseudo-Labeling: An Uncertainty-Aware Pseudo-label Selection Framework for Semi-Supervised Learningcited by 301
  3. 2018Context Encoding for Semantic Segmentationcited by 289
  4. 2019On The Power of Curriculum Learning in Training Deep Networkscited by 285
  5. 2017Automated Curriculum Learning for Neural Networkscited by 251
  6. 2015Semi-Supervised Learning with Ladder Networkscited by 246
  7. 2014Learning with Pseudo-Ensemblescited by 241
  8. 2020Rectifying Pseudo Label Learning via Uncertainty Estimation for Domain Adaptive Semantic Segmentationcited by 222
  9. 2018Curriculum Learning by Transfer Learning: Theory and Experiments with Deep Networkscited by 113
  10. 2020Multi-Task Curriculum Framework for Open-Set Semi-Supervised Learningcited by 82
  11. 2014Curriculum Learning of Multiple Taskscited by 65
  12. 2019Curriculum semi-supervised segmentationcited by 27

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2021: 4; 2022: 73; 2023: 163; 2024: 143; 2025: 97; 2026: 57.

537 papers.

Sort:

By area:

  1. 2022FreeMatch: Self-adaptive Thresholding for Semi-supervised Learningcited by 232
  2. 2022Revisiting Weak-to-Strong Consistency in Semi-Supervised Semantic Segmentationcited by 177
  3. 2023SoftMatch: Addressing the Quantity-Quality Trade-off in Semi-supervised Learningcited by 133
  4. 2022Active Learning on a Budget: Opposite Strategies Suit High and Low Budgetscited by 131
  5. 2022USB: A Unified Semi-supervised Learning Benchmark for Classificationcited by 95
  6. 2023PromptMRG: Diagnosis-Driven Prompts for Medical Report Generationcited by 92
  7. 2022Self-Training: A Surveycited by 85
  8. 2023Curricular Contrastive Regularization for Physics-aware Single Image Dehazingcited by 83
  9. 2022Debiased Self-Training for Semi-Supervised Learningcited by 62
  10. 2022Active Learning Through a Covering Lenscited by 62
  11. 2024UniMatch V2: Pushing the Limit of Semi-Supervised Semantic Segmentationcited by 53
  12. 2022PseCo: Pseudo Labeling and Consistency Training for Semi-Supervised Object Detectioncited by 51
38 more
  1. 2022Semi-supervised Semantic Segmentation with Prototype-based Consistency Regularizationcited by 50
  2. 2023CorrMatch: Label Propagation via Correlation Matching for Semi-Supervised Semantic Segmentationcited by 43
  3. 2023SemiVL: Semi-Supervised Semantic Segmentation with Vision-Language Guidancecited by 41
  4. 2023Diffusion Models and Semi-Supervised Learners Benefit Mutually with Few Labelscited by 37
  5. 2023Continual Test-time Domain Adaptation via Dynamic Sample Selectioncited by 35
  6. 2024AllSpark: Reborn Labeled Features from Unlabeled in Transformer for Semi-Supervised Semantic Segmentationcited by 35
  7. 2021Revisiting Consistency Regularization for Semi-Supervised Learningcited by 34
  8. 2023Semi-Supervised Semantic Segmentation via Gentle Teaching Assistantcited by 34
  9. 2022Towards Semi-Supervised Deep Facial Expression Recognition with An Adaptive Confidence Margincited by 32
  10. 2021Novel Class Discovery in Semantic Segmentationcited by 31
  11. 2022Memory-Guided Multi-View Multi-Domain Fake News Detectioncited by 31
  12. 2024A Survey on Data Augmentation in Large Model Eracited by 31
  13. 2023Enhancing CLIP with CLIP: Exploring Pseudolabeling for Limited-Label Prompt Tuningcited by 30
  14. 2023Logic-induced Diagnostic Reasoning for Semi-supervised Semantic Segmentationcited by 29
  15. 2023Contrastive Pseudo Learning for Open-World DeepFake Attributioncited by 27
  16. 2023Understanding and Mitigating the Label Noise in Pre-training on Downstream Taskscited by 27
  17. 2023Boosting Semi-Supervised Learning by Exploiting All Unlabeled Datacited by 26
  18. 2023CiPR: An Efficient Framework with Cross-instance Positive Relations for Generalized Category Discoverycited by 26
  19. 2022NP-Match: When Neural Processes meet Semi-Supervised Learningcited by 25
  20. 2022ConMatch: Semi-Supervised Learning with Confidence-Guided Consistency Regularizationcited by 25
  21. 2022CLNode: Curriculum Learning for Node Classificationcited by 24
  22. 2023Co-training with High-Confidence Pseudo Labels for Semi-supervised Medical Image Segmentationcited by 24
  23. 2023Exploring Structured Semantic Prior for Multi Label Recognition with Incomplete Labelscited by 24
  24. 2023Noise-Robust Fine-Tuning of Pretrained Language Models via External Guidancecited by 24
  25. 2022Mix-Teaching: A Simple, Unified and Effective Semi-Supervised Learning Framework for Monocular 3D Object Detectioncited by 23
  26. 2023Don't Stop Pretraining? Make Prompt-based Fine-tuning Powerful Learnercited by 23
  27. 2023IOMatch: Simplifying Open-Set Semi-Supervised Learning with Joint Inliers and Outliers Utilizationcited by 23
  28. 2024Curriculum Point Prompting for Weakly-Supervised Referring Image Segmentationcited by 23
  29. 2022Multiple Instance Learning via Iterative Self-Paced Supervised Contrastive Learningcited by 22
  30. 2025LEO-Split: A Semi-Supervised Split Learning Framework over LEO Satellite Networkscited by 22
  31. 2022DetMatch: Two Teachers are Better Than One for Joint 2D and 3D Semi-Supervised Object Detectioncited by 20
  32. 2023On the Effectiveness of Out-of-Distribution Data in Self-Supervised Long-Tail Learningcited by 20
  33. 2023Semi-supervised learning made simple with self-supervised clusteringcited by 20
  34. 2023LabelBench: A Comprehensive Framework for Benchmarking Adaptive Label-Efficient Learningcited by 20
  35. 2023FlatMatch: Bridging Labeled Data and Unlabeled Data with Cross-Sharpness for Semi-Supervised Learningcited by 20
  36. 2024Robust Semi-supervised Learning by Wisely Leveraging Open-set Datacited by 20
  37. 2024IC3M: In-Car Multimodal Multi-object Monitoring for Abnormal Status of Both Driver and Passengerscited by 20
  38. 2022Fuzzy Positive Learning for Semi-supervised Semantic Segmentationcited by 19

50 most cited of 537 All 537 (JSON)