Research map / Label-efficient and robust learning / Semi-supervised learning
Towards Semi-supervised Learning with Non-random Missing Labels
Yue Duan, Zhen Zhao, Lei Qi, Luping Zhou, Lei Wang, Yinghuan Shi
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| add class transition tracking Pseudo-Rectifying Guidance to FixMatch | FixMatch · CIFAR-10 MNAR gamma=20 | accuracy: 56.26 -> 94.04 | Improved |
| add PRG to SimMatch | SimMatch · CIFAR-100 MNAR gamma=50 | accuracy: 60.06 -> 65.65 | Improved |
| replace CADR with PRG on CIFAR-10 gamma=20 | FixMatch · CIFAR-10 MNAR gamma=20 | accuracy: 79.63 -> 94.04 | Improved |
| replace CADR with PRG on CIFAR-100 | FixMatch · CIFAR-100 MNAR gamma=50 | accuracy: 59.53 -> 59.11 | No clear effect |
| add PRG to FixMatch with only 40 labels and mild imbalance | FixMatch · CIFAR-10 (n_L=40, N_1=20) | accuracy: 76.53 -> 77.44 | No clear effect |
| add class-distribution re-weighting of the transition matrix (ablation) | PRG · CIFAR-10 MNAR gamma=20 | accuracy: 88.97 -> 94.04 | Improved |
| use multi-step transition H'^k with larger k instead of one step (ablation) | PRG · CIFAR-10 MNAR gamma=20 | accuracy: 94.04 -> 82.60 (k=10) | Hurt |
| use last-epoch class prediction to guide transitions (PRG Last) instead of current (ablation) | PRG · CIFAR-10 MNAR gamma=20 | accuracy: 94.04 -> 93.81 | No clear effect |
| add PRG to FixMatch in mini-ImageNet MNAR | FixMatch · mini-ImageNet gamma=50 | accuracy: 23.56 -> 44.28 | Improved |
| add PRG to VIME on tabular MNIST | VIME · Tabular MNIST | accuracy: 63.38 -> 59.41 (gamma=20); 64.80 -> 66.60 (gamma=100) | Mixed |
| use too small tracked batch number N_B (ablation) | PRG · CIFAR-10 MNAR gamma=20 | accuracy | Hurt |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 2 of 2 on most shared results.
- FixMatch — FixMatch: Simplifying Semi-Supervised Learning with Consistency and Confidence (NeurIPS 2020) · better on 5 of 5 shared results
- VIME — VIME: Extending the Success of Self- and Semi-supervised Learning to Tabular Domain (NeurIPS 2020) · better on 1 of 1 shared result
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- CGMatch: A Different Perspective of Semi-supervised Learning (CVPR 2025 · similarity 0.89)
- Keep It on a Leash: Controllable Pseudo-label Generation Towards Realistic Long-Tailed Semi-Supervised Learning (NeurIPS 2025 · similarity 0.89)
- Erasing the Bias: Fine-Tuning Foundation Models for Semi-Supervised Learning (ICML 2024 · similarity 0.88)
- Learning with Partial-Label and Unlabeled Data: A Uniform Treatment for Supervision Redundancy and Insufficiency (ICML 2024 · similarity 0.87)
- OwMatch: Conditional Self-Labeling with Consistency for Open-World Semi-Supervised Learning (NeurIPS 2024 · similarity 0.87)
- Revisiting Semi-Supervised Learning in the Era of Foundation Models (NeurIPS 2025 · similarity 0.87)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
