Research map / Label-efficient and robust learning / Semi-supervised learning
IOMatch: Simplifying Open-Set Semi-Supervised Learning with Joint Inliers and Outliers Utilization
Zekun Li, Lei Qi, Yinghuan Shi, Yang Gao
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 |
|---|---|---|---|
| use joint inlier and outlier utilization via open-set classifier loss L_op | IOMatch · CIFAR (ablation A1-A3) | closed-set accuracy | Improved |
| add multi-binary classifier alone to FixMatch-like baseline | IOMatch · CIFAR (ablation A2 vs A1) | closed-set accuracy | Improved |
| keep unsupervised training loss L_ui of closed-set classifier | IOMatch · CIFAR (ablation A3) | closed-set accuracy | Improved |
| decouple feature spaces by using a projection head for multi-binary and open-set classifiers (ablation) | IOMatch · CIFAR-50-200 | accuracy: about 2.2% drop when shared | Improved |
| add rotation recognition auxiliary objective | IOMatch · CIFAR-100 50/50, 4 labels | accuracy: 56.14 -> 58.92 | Improved |
| add SimMatch contrastive objective | IOMatch · CIFAR-100 50/50, 4 labels | accuracy: 56.14 -> 57.08 | Improved |
| lower the open-set confidence threshold tau_q (at most 0.75) | IOMatch · CIFAR | accuracy | Improved |
| vary loss weights away from 1 (too small or too large) | IOMatch · CIFAR | accuracy | Hurt |
| set tau_p to 0 versus 0.95 | IOMatch · CIFAR | accuracy | Mixed |
| present outliers in unlabeled data versus standard SSL with no outliers | IOMatch · CIFAR-50-200 | accuracy: 55.94 -> 56.14 | Improved |
| IOMatch versus strongest rivals when labels are scarce and class mismatch severe | IOMatch · CIFAR-6-24, CIFAR-20-80, CIFAR-50-200 | accuracy: +3.60%, +7.46%, +4.78% | Improved |
| IOMatch versus SimMatch when more labels are available | IOMatch · CIFAR-100 80/20, 25 labels | accuracy: 65.68 -> 64.75 | Hurt |
Later papers that compared with this method
7 later accepted papers report IOMatch as a baseline in their results tables; 6 of them report a better number on most of the results they share with it.
- Geometric Gradient Rectification for Safe Open-Set Semi-Supervised Learning (ECCV 2026 · better on 6 of 6 shared results)
- Let the Void Be Void: Robust Open-Set Semi-Supervised Learning via Selective Non-Alignment (AAAI 2026 · better on 1 of 2 shared results)
- Unsupervised Learning for Class Distribution Mismatch (ICML 2025 · better on 2 of 2 shared results)
- CaliMatch: Adaptive Calibration for Improving Safe Semi-supervised Learning (ICCV 2025 · better on 1 of 1 shared result)
- Binary Decomposition: A Problem Transformation Perspective for Open-Set Semi-Supervised Learning (ICML 2024 · better on 7 of 7 shared results)
- Robust Semi-Supervised Learning by Wisely Leveraging Open-Set Data (TPAMI 2024 · better on 3 of 3 shared results)
- SCOMatch: Alleviating Overtrusting in Open-set Semi-supervised Learning (ECCV 2024 · better on 1 of 1 shared result)
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 3 of 3 on most shared results.
- FixMatch — FixMatch: Simplifying Semi-Supervised Learning with Consistency and Confidence (NeurIPS 2020) · better on 3 of 3 shared results
- OpenMatch — OpenMatch: Open-Set Semi-supervised Learning with Open-set Consistency Regularization (NeurIPS 2021) · better on 2 of 2 shared results
- CoMatch — CoMatch: Semi-Supervised Learning With Contrastive Graph Regularization (ICCV 2021) · 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
- OpenMatch: Open-Set Semi-supervised Learning with Open-set Consistency Regularization (NeurIPS 2021 · similarity 0.95)
- SSB: Simple but Strong Baseline for Boosting Performance of Open-Set Semi-Supervised Learning (ICCV 2023 · similarity 0.94)
- Binary Decomposition: A Problem Transformation Perspective for Open-Set Semi-Supervised Learning (ICML 2024 · similarity 0.92)
- Unknown-Aware Graph Regularization for Robust Semi-supervised Learning from Uncurated Data (AAAI 2024 · similarity 0.88)
- OwMatch: Conditional Self-Labeling with Consistency for Open-World Semi-Supervised Learning (NeurIPS 2024 · similarity 0.88)
- How Does Unlabeled Data Provably Help Out-of-Distribution Detection? (ICLR 2024 · 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.
