Research map / Label-efficient and robust learning / Domain adaptation and generalization
Active Learning for Domain Adaptation: An Energy-Based Approach
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 free-energy ranking followed by MvSM uncertainty sampling and free-energy alignment | Active domain adaptation · VisDA-2017, 5% labeling budget | accuracy: 85.2 $\pm$ 0.4 -> 88.3 $\pm$ 0.1 | Helped |
| Use free-energy ranking followed by MvSM uncertainty sampling and free-energy alignment | Active domain adaptation · Office-Home, 5% labeling budget, mean | accuracy: 72.5 -> 76.7 | Helped |
| Use free-energy ranking followed by MvSM uncertainty sampling and free-energy alignment | Active domain adaptation · Office-31, 5% labeling budget, mean | accuracy: 91.1 -> 93.2 | Helped |
| Remove free-energy sampling from target selection (ablation) | EADA · Office-Home | accuracy | Hurt |
| Remove MvSM instance-uncertainty sampling from target selection (ablation) | EADA · Office-Home | accuracy | Hurt |
| Remove free-energy alignment loss (ablation) | EADA · Office-Home | accuracy | Hurt |
| Exclude the MvSM second-stage selection by setting alpha1/alpha2 to 100/1 (ablation) | EADA · VisDA-2017, Office-Home, Office-31 | accuracy: 88.3 -> 86.6; 76.7 -> 74.7; 93.2 -> 90.8 | Hurt |
| Exclude free-energy candidate selection by setting alpha1/alpha2 to 1/100 (ablation) | EADA · VisDA-2017, Office-Home, Office-31 | accuracy: 88.3 -> 85.7; 76.7 -> 72.6; 93.2 -> 90.4 | Hurt |
| Set source supervised-loss weight delta_s to 0.1 instead of the default 1.0 (ablation) | EADA · VisDA-2017 | accuracy: 88.0 -> 88.7 | Helped |
| Use EADA instead of BADGE for active selection | Toy domain-shift classification · VisDA-2017 toy example | target error rate: 4.2\% -> 1.0\% | Helped |
Closest papers by meaning
- Category-Aware Active Domain Adaptation (ICML 2024 · similarity 0.90)
- Divide and Adapt: Active Domain Adaptation via Customized Learning (CVPR 2023 · similarity 0.89)
- Revisiting the Domain Shift and Sample Uncertainty in Multi-source Active Domain Transfer (CVPR 2024 · similarity 0.89)
- Prototype-Driven Active Domain Adaptation with Density Consideration (AAAI 2026 · similarity 0.89)
- Diffusion-Based Probabilistic Uncertainty Estimation for Active Domain Adaptation (NeurIPS 2023 · similarity 0.88)
- Revisiting Source-Free Domain Adaptation: Insights into Representativeness, Generalization, and Variety (CVPR 2025 · similarity 0.87)
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.