Research map / Label-efficient and robust learning / Domain adaptation and generalization
A Prototype-Oriented Framework for Unsupervised 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 |
|---|---|---|---|
| Align target features to learned class prototypes using PCT | ResNet-50 domain adaptation · Office-31 | average accuracy (%): 88.9 -> 90.0 | Helped |
| Align target features to learned class prototypes using PCT | ResNet-50 domain adaptation · Office-Home | average accuracy (%): 68.1 -> 71.8 | Helped |
| Estimate target class proportions instead of using a uniform prior | PCT · Sub-sampled target Office-31 | average accuracy (%): 85.4 -> 86.4 | Helped |
| Estimate target class proportions instead of using a uniform prior | PCT · Sub-sampled target Office-Home | average accuracy (%): 69.8 -> 70.2 | Helped |
| Estimate target class proportions instead of using a uniform prior | PCT · Sub-sampled source data | average accuracy (%) | No clear effect |
| Add target-to-prototype transport loss (ablation) | PCT · Office-31 | average accuracy (%): 88.6 -> 90.0 | Helped |
| Add prototype-to-target transport loss (ablation) | PCT · Office-31 | average accuracy (%): 84.3 -> 90.0 | Helped |
| Use probabilistic bi-directional transport instead of exact optimal transport (ablation) | PCT · Office-31 | average accuracy (%): 87.9 -> 90.0 | Helped |
| Use probabilistic bi-directional transport instead of Sinkhorn-relaxed optimal transport (ablation) | PCT · Office-31 | average accuracy (%): 88.3 -> 90.0 | Helped |
| Add stop-gradient for prototypes in transport losses (ablation) | PCT · Office-31 | average accuracy (%): 89.0 -> 90.0 | Helped |
| Use cosine dissimilarity instead of exponential inner-product cost (ablation) | PCT · Office-31 | average accuracy (%): 86.9 -> 90.0 | Helped |
Closest papers by meaning
- Sample Selection for Universal Domain Adaptation (AAAI 2021 · similarity 0.90)
- Gradient-Based Sample Selection for Black-Box Universal Domain Adaptation (AAAI 2025 · similarity 0.89)
- Revisiting Source-Free Domain Adaptation: Insights into Representativeness, Generalization, and Variety (CVPR 2025 · similarity 0.89)
- SimProF: A Simple Probabilistic Framework for Unsupervised Domain Adaptation (AAAI 2025 · similarity 0.89)
- Tackling Dimensional Collapse toward Comprehensive Universal Domain Adaptation (ICML 2025 · similarity 0.89)
- Low Category Uncertainty and High Training Potential Instance Learning 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.