Research map / Label-efficient and robust learning / Semi-supervised learning
On Non-Random Missing Labels in Semi-Supervised Learning
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 |
|---|---|---|---|
| apply CADR rather than FixMatch under class-imbalanced labeled data | FixMatch · CIFAR-10, γ=50 | accuracy (%): 65.61 -> 93.79 | Helped |
| apply CADR rather than FixMatch under class-imbalanced labeled data | FixMatch · CIFAR-100, γ=100 | accuracy (%): 48.82 -> 60.88 | Helped |
| apply CADR rather than FixMatch under class-imbalanced labeled data | FixMatch · STL-10, γ=100 | accuracy (%): 57.01 -> 76.70 | Helped |
| apply CADR rather than FixMatch under class-imbalanced labeled data | FixMatch · mini-ImageNet, γ=100 | accuracy (%): 26.57 -> 32.78 | Helped |
| add CAP to FixMatch (ablation) | FixMatch · CIFAR-10, γ=50 | accuracy (%): 65.61 -> 89.50 | Helped |
| add CAI to FixMatch (ablation) | FixMatch · CIFAR-10, γ=50 | accuracy (%): 65.61 -> 88.15 | Helped |
| add the supplementary doubly robust loss to the CAP+CAI combination (ablation) | CAP + CAI · CIFAR-10, γ=50 | accuracy (%): 89.53 -> 93.79 | Helped |
| add the supplementary doubly robust loss to the CAP+CAI combination (ablation) | CAP + CAI · CIFAR-10, γ=100 | accuracy (%): 94.10 -> 93.97 | Hurt |
| remove moving averaging by using μ=0 (ablation) | CAP · CIFAR-10, γ=50 | accuracy (%): 89.50 -> 88.39 | Hurt |
| vary moving-average momentum from μ=0.99 to μ=0.999 (ablation) | CAP · CIFAR-10, γ=50 | accuracy (%): 89.50 -> 89.35 | No clear effect |
| set CAI threshold scaling coefficient β to 0.2 rather than 0.5 (ablation) | CAI · CIFAR-10, γ=50 | accuracy (%): 88.15 -> 65.79 | Hurt |
| change CAI threshold scaling coefficient from β=0.5 to β=1.0 (ablation) | CAI · CIFAR-10, γ=50 | accuracy (%): 88.15 -> 88.11 | No clear effect |
Closest papers by meaning
- Class-Distribution-Aware Pseudo-Labeling for Semi-Supervised Multi-Label Learning (NeurIPS 2023 · similarity 0.91)
- Barely-Supervised Learning: Semi-supervised Learning with Very Few Labeled Images (AAAI 2022 · similarity 0.90)
- OwMatch: Conditional Self-Labeling with Consistency for Open-World Semi-Supervised Learning (NeurIPS 2024 · similarity 0.89)
- CGMatch: A Different Perspective of Semi-supervised Learning (CVPR 2025 · similarity 0.89)
- Re-Evaluating the Impact of Unseen-Class Unlabeled Data on Semi-Supervised Learning Model (ICLR 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)
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.