Research map / Label-efficient and robust learning / Domain adaptation and generalization
Prototypical Partial Optimal Transport for Universal Domain Adaptation
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 m-PPOT with source prototypes instead of m-POT without prototypes (ablation) | PPOT · OPDA, Office-31, Office-Home, and VisDA | H-score (%): 88.4 -> 90.4; 74.2 -> 77.1; 66.4 -> 73.8 | Helped |
| add contrastive pre-training (ablation) | PPOT · OPDA, Office-31, Office-Home, and VisDA | H-score (%): 89.4 -> 90.4; 74.3 -> 77.1; 58.1 -> 73.8 | Helped |
| add the reweighted positive entropy loss (ablation) | PPOT · OPDA, Office-31, Office-Home, and VisDA | H-score (%): 88.4 -> 90.4; 76.5 -> 77.1; 71.1 -> 73.8 | Mixed |
| add the reweighted negative entropy loss (ablation) | PPOT · OPDA, Office-31, Office-Home, and VisDA | H-score (%): 89.6 -> 90.4; 74.4 -> 77.1; 67.8 -> 73.8 | Helped |
| use transport-plan reweighting rather than unit weights (ablation) | PPOT · OPDA, Office-31, Office-Home, and VisDA | H-score (%): 86.5 -> 90.4; 74.7 -> 77.1; 69.9 -> 73.8 | Helped |
| use PPOT | UniDA · OPDA, Office-31 | H-score (%): 87.6 -> 90.4 | Helped |
| use PPOT | UniDA · OPDA, Office-Home | H-score (%): 75.6 -> 77.1 | Helped |
| use PPOT | UniDA · OPDA, VisDA | H-score (%): 56.4 -> 73.8 | Helped |
| use PPOT | UniDA · OPDA, DomainNet average | H-score (%): 52.1 -> 57.5 | Helped |
| use PPOT without assuming source-private classes are absent | PDA methods · PDA, Office-Home | H-score (%): 79.4 -> 74.3 | Hurt |
| use PPOT without assuming source-private classes are absent | PDA methods · PDA, VisDA | H-score (%): 88.8 -> 83.0 | Hurt |
| use PPOT without assuming target-private classes are absent | OSDA methods · OSDA, Office-Home | H-score (%): 69.1 -> 70.0 | Helped |
Closest papers by meaning
- Unified Optimal Transport Framework for Universal Domain Adaptation (NeurIPS 2022 · similarity 0.94)
- Theoretical Performance Guarantees for Partial Domain Adaptation via Partial Optimal Transport (ICML 2025 · similarity 0.92)
- Tackling Dimensional Collapse toward Comprehensive Universal Domain Adaptation (ICML 2025 · similarity 0.90)
- Gradient-Based Sample Selection for Black-Box Universal Domain Adaptation (AAAI 2025 · similarity 0.90)
- Universal Semi-Supervised Domain Adaptation by Mitigating Common-Class Bias (CVPR 2024 · similarity 0.89)
- Probability-Polarized Optimal Transport for Unsupervised Domain Adaptation (AAAI 2024 · similarity 0.89)
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.