Research map / Label-efficient and robust learning / Semi-supervised learning
AdaMatch: A Unified Approach to Semi-Supervised Learning and Domain Adaptation
David Berthelot, Rebecca Roelofs, Kihyuk Sohn, Nicholas Carlini, Alexey Kurakin
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 AdaMatch instead of FixMatch+ | DigitFive UDA · DigitFive UDA | target accuracy: 95.6 -> 97.8 | Improved |
| Use AdaMatch instead of NoisyStudent | DomainNet64 UDA · DomainNet64 UDA | target accuracy: 21.4 -> 26.1 | Improved |
| Use AdaMatch instead of NoisyStudent | DomainNet224 UDA · DomainNet224 UDA | target accuracy: 23.9 -> 28.7 | Improved |
| Use ImageNet-pretrained AdaMatch rather than pretrained MCD after 8M images | DomainNet224 UDA · DomainNet224 UDA, ImageNet pre-training, 8M | target accuracy: 22.3 -> 28.2 | Improved |
| Early-stop pretrained AdaMatch at 2M images rather than train for 8M | AdaMatch · DomainNet224 UDA with ImageNet pre-training | target accuracy: 28.2 -> 33.4 | Improved |
| Use AdaMatch instead of FixMatch+ | DomainNet224 SSL with 25 labels · DomainNet224 SSL, 25 labels | target accuracy: 42.9 -> 46.4 | Improved |
| Use AdaMatch instead of FixMatch+ | DomainNet224 SSDA with 1 target label · DomainNet224 SSDA, 1 target label | target accuracy: 28.3 -> 34.8 | Improved |
| Use AdaMatch instead of FixMatch+ | DomainNet224 SSDA with 5 target labels · DomainNet224 SSDA, 5 target labels | target accuracy: 38.7 -> 42.3 | Improved |
| Add random logit interpolation (ablation) | AdaMatch · DomainNet64 UDA, 6 protocols | accuracy: 25.2 -> 26.3 | Improved |
| Add distribution alignment (ablation) | AdaMatch · DomainNet64 UDA, 6 protocols | accuracy: 17.7 -> 26.3 | Improved |
| Add relative confidence thresholding (ablation) | AdaMatch · DomainNet64 UDA, 6 protocols | accuracy: 23.2 -> 26.3 | Improved |
| Increase unlabeled-to-labeled batch ratio (ablation) | AdaMatch · DomainNet64 UDA, 6 protocols | accuracy | Improved |
Later papers that compared with this method
4 later accepted papers report AdaMatch as a baseline in their results tables; 3 of them report a better number on most of the results they share with it.
- CustomKD: Customizing Large Vision Foundation for Edge Model Improvement via Knowledge Distillation (CVPR 2025 · better on 1 of 1 shared result)
- Universal Semi-Supervised Domain Adaptation by Mitigating Common-Class Bias (CVPR 2024 · better on 2 of 2 shared results)
- Imprecise Label Learning: A Unified Framework for Learning with Various Imprecise Label Configurations (NeurIPS 2024 · better on 0 of 1 shared result)
- InstanT: Semi-supervised Learning with Instance-dependent Thresholds (NeurIPS 2023 · better on 12 of 16 shared results)
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.
- FixMatch+ — FixMatch: Simplifying Semi-Supervised Learning with Consistency and Confidence (NeurIPS 2020) · 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
- Universal Semi-Supervised Domain Adaptation by Mitigating Common-Class Bias (CVPR 2024 · similarity 0.91)
- Enhancing Semi-supervised Domain Adaptation via Effective Target Labeling (AAAI 2024 · similarity 0.89)
- Category-Aware Active Domain Adaptation (ICML 2024 · similarity 0.89)
- Tackling Dimensional Collapse toward Comprehensive Universal Domain Adaptation (ICML 2025 · similarity 0.89)
- Gradient-Based Sample Selection for Black-Box Universal Domain Adaptation (AAAI 2025 · similarity 0.88)
- A Unified Framework for Heterogeneous Semi-supervised Learning (CVPR 2025 · similarity 0.88)
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.
