Research map / Label-efficient and robust learning / Domain adaptation and generalization
Adversarial Robustness for Unsupervised Domain Adaptation
Muhammad Awais, Fengwei Zhou, Hang Xu, Lanqing Hong, Ping Luo, Sung-Ho Bae, Zhenguo Li
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 |
|---|---|---|---|
| align UDA features with robust teachers (RFA) instead of normal MDD | MDD · VisDA-2017 | PGD-20 robust accuracy (%): 4.03 -> 47.66 | Improved |
| add RFA instead of robust pre-training as backbone | MDD · VisDA-2017 | clean accuracy (%): 67.72 -> 72.90 | Improved |
| replace normal backbone with adversarially pre-trained backbone | CDAN · VisDA-2017 | clean accuracy (%): 74.23 -> 68.00 | Hurt |
| replace normal backbone with adversarially pre-trained backbone | CDAN · VisDA-2017 | PGD-20 robust accuracy (%): 0 -> 41.67 | Improved |
| add RFA to CDAN | CDAN · VisDA-2017 | PGD-20 robust accuracy (%): 0 -> 43.49 | Improved |
| use weak adversarially pre-trained student instead of normal student (ablation) | RFA · VisDA-2017 | PGD-20 robust accuracy (%): 7.63 -> 40.61 | Improved |
| use similarity preserving loss instead of L2 pairwise loss (ablation) | RFA with MDD · VisDA-2017 | PGD-20 robust accuracy (%): 8.29 -> 47.66 | Improved |
| use multiple teachers instead of a single teacher (ablation) | RFA · VisDA-2017 | clean accuracy (%): 70.31 -> 73.45 | Improved |
| use multiple teachers instead of a single teacher (ablation) | RFA · VisDA-2017 | PGD-20 robust accuracy (%): 40.15 -> 40.87 | Improved |
| increase alpha from 100 to 1000 (ablation) | RFA · VisDA-2017 | PGD-20 robust accuracy (%): 40.07 -> 47.66 | Improved |
| increase alpha from 500 to 5000 (ablation) | RFA · VisDA-2017 | clean accuracy (%): 73.62 -> 70.31 | Hurt |
| apply RFA to Office-Home | MDD · Office-Home | clean accuracy (%): 67.91 -> 70.03 | Improved |
Closest papers by meaning
- Unsupervised Robust Domain Adaptation: Paradigm, Theory and Algorithm (IJCV 2025 · similarity 0.94)
- Toward Improving Robustness and Accuracy in Unsupervised Domain Adaptation (AAAI 2025 · similarity 0.92)
- Divergence-Agnostic Unsupervised Domain Adaptation by Adversarial Attacks (TPAMI 2021 · similarity 0.92)
- Can We Evaluate Domain Adaptation Models Without Target-Domain Labels? (ICLR 2024 · similarity 0.92)
- Adversarial Unsupervised Domain Adaptation With Conditional and Label Shift: Infer, Align and Iterate (ICCV 2021 · similarity 0.92)
- CASUAL: Conditional Support Alignment for Domain Adaptation with Label Shift (AAAI 2025 · 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.
