Research map / Label-efficient and robust learning / Domain adaptation and generalization
Open Set Domain Adaptation
Pau Panareda Busto, Juergen Gall
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 |
|---|---|---|---|
| apply assign-and-transform-iteratively (ATI) domain adaptation instead of no adaptation | linear SVM · Office unsupervised open set (OS average) | accuracy (%): 68.3 -> 82.6 | Improved |
| apply ATI instead of CNN-based adaptation (BP) | ATI · Office unsupervised open set (OS average) | accuracy (%): 77.4 -> 82.6 | Improved |
| add outlier rejection (lambda) to ATI (ablation) | ATI · Office unsupervised open set, subset of samples (OS average) | accuracy (%): 81.5 -> 82.3 | Improved |
| add outlier rejection (lambda) to ATI (ablation) | ATI · Office unsupervised, all samples (CS average) | accuracy (%): 93.8 -> 93.7 | No clear effect |
| add locality constraint with 1 neighbour in unsupervised setting (ablation) | ATI-λ · Office unsupervised open set (OS average) | accuracy (%): 82.9 -> 81.3 | Hurt |
| add locality constraint with 1 neighbour in semi-supervised setting (ablation) | ATI-λ · Office semi-supervised open set (OS average) | accuracy (%): 85.7 -> 86.0 | Improved |
| use ATI instead of CNN method MMD in semi-supervised setting | ATI · Office semi-supervised open set (OS average) | accuracy (%): 80.9 -> 85.7 | Improved |
| include the unknown class in the label set C for assignment (ablation) | ATI-λ · Office unsupervised open set (OS average) | accuracy (%): 82.3 -> 82.9 | Improved |
| train with negative unknown samples (standard linear SVM) instead of open set SVM (ablation) | ATI-λ · Office unsupervised open set (OS average) | accuracy (%): 75.0 -> 82.9 | Improved |
| increase unknown samples from 20 to 400 in source and target (ablation) | LSVM · Amazon -> DSLR+Webcam open set | accuracy (%): 74.2 -> 48.2 | Hurt |
| increase unknown samples from 20 to 400 in source and target (ablation) | ATI-λ · Amazon -> DSLR+Webcam open set | accuracy (%): 80.3 -> 81.9 | No clear effect |
| apply ATI-λ instead of CORAL on dense cross-dataset | ATI-λ · Dense cross-dataset open set (average) | accuracy (%): 44.2 -> 50.2 | Improved |
Closest papers by meaning
- Open Set Domain Adaptation for Image and Action Recognition (TPAMI 2018 · similarity 0.93)
- Separate to Adapt: Open Set Domain Adaptation via Progressive Separation (CVPR 2019 · similarity 0.92)
- Open Set Domain Adaptation by Backpropagation (ECCV 2018 · similarity 0.91)
- Open-Set Graph Domain Adaptation via Separate Domain Alignment (AAAI 2024 · similarity 0.87)
- Open-Set Heterogeneous Domain Adaptation: Theoretical Analysis and Algorithm (AAAI 2025 · similarity 0.87)
- Distinguish Then Exploit: Source-free Open Set Domain Adaptation via Weight Barcode Estimation and Sparse Label Assignment (CVPR 2025 · similarity 0.86)
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.
