Research map / Label-efficient and robust learning / Domain adaptation and generalization
Subsidiary Prototype Alignment 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 |
|---|---|---|---|
| add SPA (word-prototypes + pretext task + entropy loss) | OVANet · UniDA Office-Home | HOS: 71.8 -> 73.2 | Helped |
| add SPA | DCC · UniDA Office-Home | HOS: 70.2 -> 74.9 | Helped |
| add SPA | OVANet · UniDA DomainNet | HOS: 50.7 -> 52.2 | Helped |
| add SPA | DCC · OSDA Office-31 | HOS: 86.8 -> 93.2 | Helped |
| add architecture modification only (ablation) | OVANet · UniDA Office-Home | HOS: 71.8 -> 71.8 | No clear effect |
| add architecture modification only (ablation) | DCC · UniDA Office-Home | HOS: 70.2 -> 70.9 | Helped |
| add word-histogram self-entropy loss to arch mod (ablation) | OVANet · UniDA Office-Home | HOS: 71.8 -> 72.2 | Helped |
| add pretext task without entropy loss (ablation) | DCC · UniDA Office-Home | HOS: 70.2 -> 73.6 | Helped |
| add entropy loss on top of pretext task (full SPA) (ablation) | DCC · UniDA Office-Home | HOS: 73.6 -> 74.9 | Helped |
| add entropy loss on top of pretext task (full SPA) (ablation) | OVANet · UniDA Office-Home | HOS: 72.8 -> 73.2 | Helped |
| use rotation pretext task instead of proposed grid-shuffle instance-count task (ablation) | OVANet+arch-mod · UniDA Office-Home | HOS: 73.2 -> 72.4 | Hurt |
| use colorization pretext task instead of proposed one (ablation) | OVANet+arch-mod · UniDA Office-Home | HOS: 73.2 -> 71.8 | Hurt |
Closest papers by meaning
- Target Semantics Clustering via Text Representations for Robust Universal Domain Adaptation (AAAI 2025 · similarity 0.90)
- Prompt-Based Distribution Alignment for Unsupervised Domain Adaptation (AAAI 2024 · similarity 0.90)
- Unified Language-driven Zero-shot Domain Adaptation (CVPR 2024 · similarity 0.89)
- Tackling Dimensional Collapse toward Comprehensive Universal Domain Adaptation (ICML 2025 · similarity 0.89)
- Universal Domain Adaptation for Semantic Segmentation (CVPR 2025 · similarity 0.89)
- Memory-Assisted Sub-Prototype Mining for Universal Domain Adaptation (ICLR 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.