Research map / Label-efficient and robust learning / Domain adaptation and generalization
Keypoint-Guided Optimal Transport with Applications in Heterogeneous Domain Adaptation
Xiang Gu, Yucheng Yang, Wei Zeng, Jian Sun, Zongben Xu
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 relation-preserving guidance instead of distance-preserving guidance (ablation) | KPG-RL (SH) · Office-31 HDA | accuracy: 64.9 -> 79.9 | Improved |
| solve with Sinkhorn instead of linear programming (ablation) | KPG-RL · Office-31 HDA | accuracy: 79.3 -> 79.9 | No clear effect |
| use GW-based variant instead of Sinkhorn variant (ablation) | KPG-RL · Office-31 HDA | accuracy: 79.9 -> 79.6 | No clear effect |
| add mask-based constraint on transport plan (ablation) | GW · Office-31 HDA | accuracy: 35.6 -> 60.8 | Improved |
| add mask-based constraint on transport plan (ablation) | HOT · Office-31 HDA | accuracy: 44.3 -> 55.9 | Improved |
| add mask-based constraint on transport plan (ablation) | TLB · Office-31 HDA | accuracy: 35.7 -> 58.3 | Improved |
| plain GW/HOT/TLB transport vs labeled target only | GW · Office-31 HDA | accuracy: 60.6 -> 35.6 | Hurt |
| relation-preserving KPG-RL vs class-center-preserving SGW | KPG-RL (SH) · Office-31 HDA | accuracy: 67.3 -> 79.9 | Improved |
| add KPG-RL transport with 1 labeled target shot per class | Labeled-target-only · Office-31 HDA 1-shot | accuracy: 60.6 -> 79.9 | Improved |
| add KPG-RL transport with 3 labeled target shots per class | Labeled-target-only · Office-31 HDA 3-shot | accuracy: 80.9 -> 87.7 | Improved |
| more labeled target shots (1 to 3) shrinks gain of transfer | KPG-RL (SH) · Office-31 HDA | accuracy: 79.9 -> 87.7 | Mixed |
| add partial-KPG-RL detection of unknown classes over rejection baseline | Baseline · Office-31 open-set HDA | HOS: 48.7 -> 69.1 | Improved |
Closest papers by meaning
- Keypoint-Guided Optimal Transport: Models, Algorithms, and Applications (JMLR 2026 · similarity 0.94)
- Theoretical Performance Guarantees for Partial Domain Adaptation via Partial Optimal Transport (ICML 2025 · similarity 0.84)
- Probability-Polarized Optimal Transport for Unsupervised Domain Adaptation (AAAI 2024 · similarity 0.84)
- Optimal Flow Matching: Learning Straight Trajectories in Just One Step (NeurIPS 2024 · similarity 0.84)
- Low-Rank Optimal Transport through Factor Relaxation with Latent Coupling (NeurIPS 2024 · similarity 0.83)
- OTMatch: Improving Semi-Supervised Learning with Optimal Transport (ICML 2024 · similarity 0.83)
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.
