Research map / Label-efficient and robust learning / OOD and anomaly detection
ProSub: Probabilistic Open-Set Semi-Supervised Learning with Subspace-Based Out-of-Distribution Detection
Erik Wallin, Lennart Svensson, Fredrik Kahl, Lars Hammarstrand
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 |
|---|---|---|---|
| use subspace score instead of energy score for ID/OOD detection (ablation) | ProSub (end of warm-up) · CIFAR-100 / CIFAR-10 | AUROC: 0.65 -> 0.92 | Improved |
| use subspace score for a fully supervised model (ablation) | fully supervised model · CIFAR-100 / CIFAR-10 | AUROC: 0.81 (energy) -> 0.73 (s) | Hurt |
| replace angle score with residual distance or max similarity (ablation) | ProSub · CIFAR-100 / CIFAR-10 | AUROC: 0.92 -> 0.88 (residual), 0.87 (max sim) | Hurt |
| replace probabilistic ID/OOD with Otsu hard threshold (ablation) | ProSub · CIFAR-100 (2,500 labels) / CIFAR-10 | AUROC: 0.97 -> 0.67 | Hurt |
| use predicted probabilities directly (weighted) instead of sampled ID mask (ablation) | ProSub · CIFAR-100 (2,500 labels) / CIFAR-10 | accuracy: 74.11 -> 74.01 | No clear effect |
| add self-supervision loss l_self (ablation) | ProSub · CIFAR-100 (10,000 labels) / CIFAR-10 | AUROC: 0.63 -> 0.87 | Improved |
| add subspace loss l_sub (ablation) | ProSub · CIFAR-100 (10,000 labels) / CIFAR-10 | AUROC: 0.63 -> 0.96 | Improved |
| combine l_self and l_sub (ablation) | ProSub · CIFAR-100 (10,000 labels) / CIFAR-10 | accuracy: 70.24 -> 79.46 | Improved |
| ProSub vs OpenMatch with CIFAR-10 as ID | ProSub · CIFAR-10 ID, 1,000 labels | AUROC: 0.92 vs 0.93 | Hurt |
Later papers that compared with this method
3 later accepted papers report ProSub as a baseline in their results tables; 3 of them report a better number on most of the results they share with it.
- Let the Void Be Void: Robust Open-Set Semi-Supervised Learning via Selective Non-Alignment (AAAI 2026 · better on 4 of 4 shared results)
- PAF: Perturbation-Aware Filtering for Open-Set Semi-Supervised Learning (CVPR 2026 · better on 3 of 3 shared results)
- Experimental Design for Multi-Channel Imaging via Task-Driven Feature Selection (ICLR 2024 · better on 4 of 4 shared results)
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 4 of 4 shared results
- OpenMatch — OpenMatch: Open-Set Semi-supervised Learning with Open-set Consistency Regularization (NeurIPS 2021) · better on 2 of 3 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
- Robust Semi-Supervised Learning by Wisely Leveraging Open-Set Data (TPAMI 2024 · similarity 0.90)
- STEP: Out-of-Distribution Detection in the Presence of Limited In-Distribution Labeled Data (NeurIPS 2021 · similarity 0.89)
- Provable Discriminative Hyperspherical Embedding for Out-of-Distribution Detection (AAAI 2025 · similarity 0.86)
- EAT: Towards Long-Tailed Out-of-Distribution Detection (AAAI 2024 · similarity 0.86)
- Learning to Shape In-distribution Feature Space for Out-of-distribution Detection (NeurIPS 2024 · similarity 0.86)
- The Best of Both Worlds: On the Dilemma of Out-of-distribution Detection (NeurIPS 2024 · similarity 0.85)
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.
