Research map / Label-efficient and robust learning / Semi-supervised learning
Geometric Gradient Rectification for Safe Open-Set Semi-Supervised Learning
Jiahe Chen, Qian Shao, Qiyuan Chen, Jiaying He, Jintai Chen, Hongxia Xu, Jian Wu
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 GGR gradient rectification | IOMatch · CIFAR-10 6/4 split, 5 labels/class (closed-set) | accuracy: 89.87 -> 91.87 | Improved |
| add GGR gradient rectification | DAC · CIFAR-10 6/4 split, 5 labels/class (closed-set) | accuracy: 86.69 -> 87.63 | Improved |
| add GGR gradient rectification | OpenMatch · CIFAR-10 6/4 split, 5 labels/class (closed-set) | accuracy: 47.26 -> 47.38 | No clear effect |
| add GGR gradient rectification | IOMatch · CIFAR-100 20/80 split, 5 labels/class (closed-set) | accuracy: 56.48 -> 56.75 | No clear effect |
| add GGR gradient rectification | IOMatch · ImageNet-30 20/10, 1% labels | accuracy: 85.18 -> 85.90 | Improved |
| add GGR gradient rectification | DAC · ImageNet-30 20/10, 1% labels | accuracy: 86.40 -> 87.40 | Improved |
| add GGR gradient rectification | OpenMatch · ImageNet-30 20/10, 5% labels (open-set) | balanced accuracy: 61.51 -> 68.15 | Improved |
| add GGR gradient rectification | DAC · CIFAR-100 50/50 split, 5 labels/class (open-set BA) | balanced accuracy: 54.65 -> 53.77 | Hurt |
| apply GGR to head instead of backbone (ablation) | DAC · CIFAR-10 6/4 split, 5 labels/class (closed-set) | accuracy: 87.63 -> 90.38 | Improved |
| apply GGR to both backbone and head instead of backbone (ablation) | IOMatch · CIFAR-10 6/4 split, 5 labels/class (closed-set) | accuracy: 91.87 -> 89.56 | Hurt |
| apply GGR to both backbone and head instead of backbone (ablation) | DAC · CIFAR-100 50/50 split, 5 labels/class (closed-set) | accuracy: 57.19 -> 58.50 | Improved |
| use CSR subspace rectification instead of VLR (ablation) | DAC · CIFAR-100 50/50 (efficiency study) | accuracy: 56.90 -> 57.26 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- OpenMatch — OpenMatch: Open-Set Semi-supervised Learning with Open-set Consistency Regularization (NeurIPS 2021) · better on 4 of 4 shared results
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
- Out-Of-Domain Unlabeled Data Improves Generalization (ICLR 2024 · similarity 0.85)
- Targeted Representation Alignment for Open-World Semi-Supervised Learning (CVPR 2024 · similarity 0.85)
- CGMatch: A Different Perspective of Semi-supervised Learning (CVPR 2025 · similarity 0.84)
- Unknown-Aware Graph Regularization for Robust Semi-supervised Learning from Uncurated Data (AAAI 2024 · similarity 0.84)
- Gradient-Weight Alignment as a Train-Time Proxy for Generalization in Classification Tasks (NeurIPS 2025 · similarity 0.84)
- Binary Decomposition: A Problem Transformation Perspective for Open-Set Semi-Supervised Learning (ICML 2024 · similarity 0.84)
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.
