mlsanityData to 14 Oct 2026

SemiReward: A General Reward Model for Semi-supervised Learning

arXiv 2310.03013 ↗First posted 4 Oct 2023cs.LG cs.AIcited by 14

On the map: LLM alignment / RLHF / preference optimisation → Reward modeling and RLHF

Abstract (arXiv)

Semi-supervised learning (SSL) has witnessed great progress with various improvements in the self-training framework with pseudo labeling. The main challenge is how to distinguish high-quality pseudo labels against the confirmation bias. However, existing pseudo-label selection strategies are limited to pre-defined schemes or complex hand-crafted policies specially designed for classification, failing to achieve high-quality labels, fast convergence, and task versatility simultaneously. To these ends, we propose a Semi-supervised Reward framework (SemiReward) that predicts reward scores to evaluate and filter out high-quality pseudo labels, which is pluggable to mainstream SSL methods in wide task types and scenarios. To mitigate confirmation bias, SemiReward is trained online in two stages with a generator model and subsampling strategy. With classification and regression tasks on 13 standard SSL benchmarks across three modalities, extensive experiments verify that SemiReward achieves significant performance gains and faster convergence speeds upon Pseudo Label, FlexMatch, and Free/SoftMatch. Code and models are available at https://github.com/Westlake-AI/SemiReward.

Key facts. Cited by 14 papers in the map (data through 14 Oct 2026). Placed in Reward modeling and RLHF, one of 10 directions in LLM alignment / RLHF / preference optimisation. Builds on 57 of its 81 references that are in the map. First posted on arXiv 4 Oct 2023.

Builds on

57 of 81 references.

Closest ancestors · 3

Same direction, newest first

  1. 2022SURF: Semi-supervised Reward Learning with Data Augmentation for Feedback-efficient Preference-based Reinforcement Learning18 Mar
  2. 2018Scalable agent alignment via reward modeling: a research direction19 Nov
  3. 2017Deep reinforcement learning from human preferences12 Jun

Other references · 54

  1. 2015Deep Residual Learning for Image Recognitioncited by 52,096
  2. 2018BERT: Pre-training of Deep Bidirectional Transformers for Language Understandingcited by 40,399
  3. 2014Adam: A Method for Stochastic Optimizationcited by 37,240
  4. 2020An Image is Worth 16x16 Words: Transformers for Image Recognition at Scalecited by 30,186
  5. 2022Training language models to follow instructions with human feedbackcited by 15,971
  6. 2017Proximal Policy Optimization Algorithmscited by 15,215
  7. 2020A Simple Framework for Contrastive Learning of Visual Representationscited by 12,049
  8. 2016SGDR: Stochastic Gradient Descent with Warm Restartscited by 5,045
  9. 2021HuBERT: Self-Supervised Speech Representation Learning by Masked Prediction of Hidden Unitscited by 2,342
  10. 2015Character-level Convolutional Networks for Text Classificationcited by 2,330
  11. 2020FixMatch: Simplifying Semi-Supervised Learning with Consistency and Confidencecited by 2,085
  12. 2017EuroSAT: A Novel Dataset and Deep Learning Benchmark for Land Use and Land Cover Classificationcited by 1,496
38 more
  1. 2019MixMatch: A Holistic Approach to Semi-Supervised Learningcited by 1,391
  2. 2019Self-training with Noisy Student improves ImageNet classificationcited by 1,311
  3. 2019Unsupervised Data Augmentation for Consistency Trainingcited by 1,060
  4. 2016Temporal Ensembling for Semi-Supervised Learningcited by 1,057
  5. 2017Virtual Adversarial Training: A Regularization Method for Supervised and Semi-Supervised Learningcited by 920
  6. 2018AutoAugment: Learning Augmentation Policies from Datacited by 730
  7. 2020DivideMix: Learning with Noisy Labels as Semi-supervised Learningcited by 692
  8. 2021FlexMatch: Boosting Semi-Supervised Learning with Curriculum Pseudo Labelingcited by 540
  9. 2019Pseudo-Labeling and Confirmation Bias in Deep Semi-Supervised Learningcited by 510
  10. 2017Mean teachers are better role models: Weight-averaged consistency targets improve semi-supervised deep learning resultscited by 472
  11. 2020Meta Pseudo Labelscited by 374
  12. 2019ReMixMatch: Semi-Supervised Learning with Distribution Alignment and Augmentation Anchoringcited by 349
  13. 2021Unbiased Teacher for Semi-Supervised Object Detectioncited by 306
  14. 2015Semi-Supervised Learning with Ladder Networkscited by 247
  15. 2022FreeMatch: Self-adaptive Thresholding for Semi-supervised Learningcited by 232
  16. 2018Explicit Inductive Bias for Transfer Learning with Convolutional Networkscited by 227
  17. 2020Mutual Mean-Teaching: Pseudo Label Refinery for Unsupervised Domain Adaptation on Person Re-identificationcited by 215
  18. 2020CoMatch: Semi-supervised Learning with Contrastive Graph Regularizationcited by 167
  19. 2021Dash: Semi-Supervised Learning with Dynamic Thresholdingcited by 148
  20. 2021All Tokens Matter: Token Labeling for Training Better Vision Transformerscited by 144
  21. 2023SoftMatch: Addressing the Quantity-Quality Trade-off in Semi-supervised Learningcited by 133
  22. 2022SimMatch: Semi-supervised Learning with Similarity Matchingcited by 118
  23. 2021AdaMatch: A Unified Approach to Semi-Supervised Learning and Domain Adaptationcited by 104
  24. 2016Semi-Supervised Learning with Generative Adversarial Networkscited by 103
  25. 2022USB: A Unified Semi-supervised Learning Benchmark for Classificationcited by 95
  26. 2017Good Semi-supervised Learning that Requires a Bad GANcited by 86
  27. 2019DELTA: DEep Learning Transfer using Feature Map with Attention for Convolutional Networkscited by 80
  28. 2022Debiased Self-Training for Semi-Supervised Learningcited by 62
  29. 2022C-Mixup: Improving Generalization in Regressioncited by 57
  30. 2021Localization Distillation for Dense Object Detectioncited by 55
  31. 2021Co-learning: Learning from Noisy Labels with Self-supervisioncited by 51
  32. 2017Adversarial Dropout for Supervised and Semi-supervised Learningcited by 39
  33. 2018Strong Baselines for Neural Semi-supervised Learning under Domain Shiftcited by 39
  34. 2021Self-Tuning for Data-Efficient Deep Learningcited by 39
  35. 2021Revisiting Consistency Regularization for Semi-Supervised Learningcited by 34
  36. 2022ConMatch: Semi-Supervised Learning with Confidence-Guided Consistency Regularizationcited by 25
  37. 2019Learning Sound Event Classifiers from Web Audio with Noisy Labelscited by 22
  38. 2021The Semi-Supervised iNaturalist-Aves Challenge at FGVC7 Workshopcited by 21

50 most cited of 54 All 54 (JSON)

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2023: 1; 2024: 5; 2025: 5; 2026: 3.

14 papers.

Sort:

By area:

  1. 2024Retrieval Meets Reasoning: Even High-school Textbook Knowledge Benefits Multimodal Reasoningcited by 14
  2. 2024SUMix: Mixup with Semantic and Uncertain Informationcited by 12
  3. 2024Switch EMA: A Free Lunch for Better Flatness and Sharpnesscited by 9
  4. 2024AlchemistCoder: Harmonizing and Eliciting Code Capability by Hindsight Tuning on Multi-source Datacited by 6
  5. 2025When Confidence Fails: Revisiting Pseudo-Label Selection in Semi-supervised Semantic Segmentationcited by 6
  6. 2025PacGDC: Label-Efficient Generalizable Depth Completion with Projection Ambiguity and Consistencycited by 4
  7. 2025OmniGaze: Reward-inspired Generalizable Gaze Estimation In The Wildcited by 4
  8. 2023A Mixture of Exemplars Approach for Efficient Out-of-Distribution Detection with Foundation Modelscited by 2
  9. 2025Keep It on a Leash: Controllable Pseudo-label Generation Towards Realistic Long-Tailed Semi-Supervised Learningcited by 1
  10. 2026PaperFit: Vision-in-the-Loop Typesetting Optimization for Scientific Documentscited by 1
  11. 2024LiqD: A Dynamic Liquid Level Detection Model under Tricky Small Containerscited by 0
  12. 2025Learning Adaptive Pseudo-Label Selection for Semi-Supervised 3D Object Detectioncited by 0
2 more
  1. 2026CoVar: Confidence-Variance-Guided Pseudo-Label Selection for Semi-Supervised Learningcited by 0
  2. 2026Beyond Distribution Estimation: Simplex Anchored Structural Inference Towards Universal Semi-Supervised Learningcited by 0

Similar, not cited

Close in wording, no citation either way

  1. 2024Semi-Supervised Reward Modeling via Iterative Self-Training
  2. 2024Reinforcement Learning-Guided Semi-Supervised Learning
  3. 2025Unlabeled Data vs. Pre-trained Knowledge: Rethinking SSL in the Era of Large Models
  4. 2025Semi-Supervised Preference Optimization with Limited Feedback
  5. 2025Enhancing Semi-supervised Learning with Zero-shot Pseudolabels
  6. 2025Toward Robust Semi-supervised Regression via Dual-stream Knowledge Distillation
  7. 2023Diffusion Models and Semi-Supervised Learners Benefit Mutually with Few Labels
  8. 2023Prompt-based Pseudo-labeling Strategy for Sample-Efficient Semi-Supervised Extractive Summarization