Research map / Label-efficient and robust learning / Domain adaptation and generalization
A Balanced and Uncertainty-aware Approach for Partial Domain Adaptation
Jian Liang, Yunbo Wang, Dapeng Hu, Ran He, Jiashi Feng
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 |
|---|---|---|---|
| add balanced adversarial alignment (target augmentation with source samples) to E-DANN (ablation) | E-DANN · Office-Home partial DA | accuracy (avg): 70.24 -> 73.40 | Improved |
| add balanced adversarial alignment to E-DANN (ablation) | E-DANN · Office31 partial DA | accuracy (avg): 95.60 -> 96.20 | Improved |
| add balanced adversarial alignment to E-DANN (ablation) | E-DANN · ImageNet-Caltech partial DA | accuracy (avg): 78.00 -> 81.16 | Improved |
| add adaptive complement entropy (uncertainty suppression) to BAA (ablation) | Ours (w/ BAA) · Office-Home partial DA | accuracy (avg): 73.40 -> 75.98 | Improved |
| add adaptive complement entropy to BAA (ablation) | Ours (w/ BAA) · Office31 partial DA | accuracy (avg): 96.20 -> 97.81 | Improved |
| add adaptive complement entropy to BAA (ablation) | Ours (w/ BAA) · ImageNet-Caltech partial DA | accuracy (avg): 81.16 -> 83.68 | Improved |
| add uncertainty suppression to closed-set DA (ablation) | Ours (w/ BAA) · Office-Home closed-set DA | accuracy (avg): 67.6 -> 68.7 | Improved |
| add uncertainty suppression to closed-set DA with CDAN (ablation) | Ours (w/ BAA)+CDAN · Office-Home closed-set DA | accuracy (avg): 68.0 -> 69.2 | Improved |
| add balanced adversarial alignment to closed-set DA (ablation) | DANN-based · Office-Home closed-set DA | accuracy (avg) | No clear effect |
| use adaptive complement entropy (xi=0.3) instead of plain complement entropy (xi=0) (ablation) | BA3US · Office-Home partial DA | accuracy (avg): 75.32 -> 76.28 | Improved |
| use adaptive complement entropy (xi=0.9) instead of xi=0 (ablation) | BA3US · Office31 partial DA | accuracy (avg): 97.68 -> 97.84 | No clear effect |
| set beta=10.0 instead of 1.0 (ablation) | BA3US · Office-Home partial DA | accuracy (avg): 75.98 -> 73.25 | Hurt |
Closest papers by meaning
- Implicit Class-Conditioned Domain Alignment for Unsupervised Domain Adaptation (ICML 2020 · similarity 0.93)
- Domain Adaptation with Asymmetrically-Relaxed Distribution Alignment (ICML 2019 · similarity 0.92)
- Partial Adversarial Domain Adaptation (ECCV 2018 · similarity 0.92)
- Minimum Class Confusion for Versatile Domain Adaptation (ECCV 2020 · similarity 0.91)
- CASUAL: Conditional Support Alignment for Domain Adaptation with Label Shift (AAAI 2025 · similarity 0.90)
- Style Adaptation and Uncertainty Estimation for Multi-Source Blended-Target Domain Adaptation (NeurIPS 2024 · similarity 0.90)
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.
