Research map / Label-efficient and robust learning / Domain adaptation and generalization
Optimal Transport for Domain Adaptation
Nicolas Courty, Rémi Flamary, Devis Tuia, Alain Rakotomamonjy
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 entropic regularization to exact OT (ablation) | OT-exact · Digits mean (SURF) | accuracy: 49.96 -> 59.20 | Improved |
| add class-based group-lasso regularizer to entropic OT (ablation) | OT-IT · Office-Caltech mean (SURF) | accuracy: 42.30 -> 47.70 | Improved |
| add class-based group-lasso regularizer to entropic OT (ablation) | OT-IT · Digits mean (SURF) | accuracy: 59.20 -> 63.90 | Improved |
| add Laplacian regularizer to entropic OT (ablation) | OT-IT · Faces PIE mean (SURF) | accuracy: 54.89 -> 56.10 | Improved |
| OT-GL vs JDA on PIE faces (ablation) | OT-GL · Faces PIE mean (SURF) | accuracy: 56.69 vs 55.88 | Hurt |
| use OT-GL instead of 1NN without adaptation | 1NN · Office-Caltech mean (SURF) | accuracy: 28.47 -> 47.70 | Improved |
| add OT-GL adaptation to DeCAF layer 6 features | DeCAF · Office-Caltech mean | accuracy: 65.20 -> 88.18 | Improved |
| use layer 7 instead of layer 6 features with OT-GL (ablation) | OT-GL · Office-Caltech mean | accuracy: 88.18 -> 88.11 | No clear effect |
| add labeled target samples to transport plan (semi-supervised term) (ablation) | OT-GL · Office-Caltech mean (SURF) | accuracy: 46.6 -> 55.6 | Improved |
| add semi-supervised term (ablation) | OT-IT · Office-Caltech mean (SURF) | accuracy: 41.8 -> 54.8 | Improved |
| OT-GL semi-supervised vs MMDT | OT-GL · Office-Caltech mean (SURF) | accuracy: 52.5 -> 55.6 | Improved |
| add Laplacian regularization to OT at 40 degree rotation (ablation) | OT-exact · Two moons | error rate: 0.109 -> 0.062 | Improved |
Closest papers by meaning
- Joint distribution optimal transportation for domain adaptation (NeurIPS 2017 · similarity 0.91)
- Domain Adaptation with Invariant Representation Learning: What Transformations to Learn? (NeurIPS 2021 · similarity 0.91)
- Reliable Weighted Optimal Transport for Unsupervised Domain Adaptation (CVPR 2020 · similarity 0.90)
- Enhanced Transport Distance for Unsupervised Domain Adaptation (CVPR 2020 · similarity 0.89)
- Theoretical Performance Guarantees for Partial Domain Adaptation via Partial Optimal Transport (ICML 2025 · similarity 0.89)
- Probability-Polarized Optimal Transport for Unsupervised Domain Adaptation (AAAI 2024 · 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.
