Research map / Label-efficient and robust learning / Domain adaptation and generalization
Homeomorphism Alignment for Unsupervised Domain Adaptation
Lihua Zhou, Mao Ye, Xiatian Zhu, Siying Xiao, Xu-Qian Fan, Ferrante Neri
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 INN homeomorphism instead of ordinary bijection with two networks (ablation) | HMA(CAN) · Office-31 A->D | accuracy (%): 90.3 -> 95.8 | Improved |
| use INN instead of two networks reduces parameters (ablation) | HMA(CAN) · Office-31 | parameters: 33603584 -> 20992000 | Improved |
| add INN mapping alone with source retraining (ablation) | CAN · Office-31 | accuracy (%): A->W 94.5 -> 94.1 | No clear effect |
| train in target space with transformed source features (S2T) (ablation) | INN(DAN) · Office-31 D->A | accuracy (%): 62.9 -> 66.5 | Improved |
| add target prediction consistency (T2S) (ablation) | INN(DAN)+S2T · Office-31 D->A | accuracy (%): 66.5 -> 67.9 | Improved |
| use class-conditional sewing (CAN) instead of marginal (DAN) matching (ablation) | HMA · Office-31 | avg accuracy (%): 83.9 -> 91.2 | Improved |
| apply HMA on top of CAN | CAN · Office-Home | avg accuracy (%): 71.8 -> 73.2 | Improved |
| apply HMA on top of DANN | DANN · Office-Home | avg accuracy (%): 57.6 -> 62.7 | Improved |
| apply HMA on top of DAN | DAN · Office-31 | avg accuracy (%): 80.4 -> 83.9 | Improved |
| HMA(CAN) vs ATDOC on Office-Home | HMA(CDAN) · Office-Home | avg accuracy (%): HMA(CDAN) 71.7 vs ATDOC 72.2 | Mixed |
Closest papers by meaning
- Partial Domain Adaptation Without Domain Alignment (TPAMI 2022 · similarity 0.90)
- Progressive Feature Alignment for Unsupervised Domain Adaptation (CVPR 2019 · similarity 0.90)
- Prior Knowledge Guided Unsupervised Domain Adaptation (ECCV 2022 · similarity 0.89)
- Co-regularized Alignment for Unsupervised Domain Adaptation (NeurIPS 2018 · similarity 0.89)
- Unsupervised Domain Adaptation With Hierarchical Gradient Synchronization (CVPR 2020 · similarity 0.88)
- Prompt-Based Distribution Alignment 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.
