Research map / Label-efficient and robust learning / Domain adaptation and generalization
Adversarial Domain Adaptation with Domain Mixup
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 DM-ADA rather than ADR | Digit adaptation · MN → US (p) | accuracy (%): 93.2 ± 2.5 -> 94.8 ± 0.7 | Helped |
| Use DM-ADA rather than ADR | Digit adaptation · MN → US (f) | accuracy (%): 96.1 ± 0.3 -> 96.7 ± 0.5 | Helped |
| Use DM-ADA rather than ADR | Digit adaptation · US → MN | accuracy (%): 93.1 ± 1.3 -> 94.2 ± 0.9 | Helped |
| Use DM-ADA rather than ADR | Digit adaptation · SV → MN | accuracy (%): 95.0 ± 1.9 -> 95.5 ± 1.1 | Helped |
| Use DM-ADA rather than GCAN | Office-31 adaptation · A → W | accuracy (%): 82.7 ± 0.1 -> 83.9 ± 0.4 | Helped |
| Add pixel-level mixup to the baseline (ablation) | DM-ADA · Office-31 A → W | accuracy (%): 76.7 -> 78.1 | Helped |
| Add feature-level mixup to the baseline (ablation) | DM-ADA · Office-31 A → W | accuracy (%): 76.7 -> 82.1 | Helped |
| Add pixel-level mixup to feature-level mixup (ablation) | DM-ADA · Office-31 A → W | accuracy (%): 82.1 -> 83.2 | Helped |
| Add triplet loss to pixel-level mixup (ablation) | DM-ADA · Office-31 A → W | accuracy (%): 78.1 -> 79.4 | Helped |
| Add triplet loss to both levels of mixup (ablation) | DM-ADA · Office-31 A → W | accuracy (%): 83.2 -> 83.9 | Helped |
| Add the discriminator class branch (ablation) | DM-ADA · Office-31 A → W | accuracy (%): 80.6 -> 82.3 | Helped |
| Add pseudo target labels to the discriminator class branch (ablation) | DM-ADA · Office-31 A → W | accuracy (%): 82.3 -> 83.9 | Helped |
Closest papers by meaning
- Adversarial Discriminative Domain Adaptation (CVPR 2017 · similarity 0.91)
- On Unsupervised Domain Adaptation: Pseudo Label Guided Mixup for Adversarial Prompt Tuning (AAAI 2024 · similarity 0.90)
- f-Domain Adversarial Learning: Theory and Algorithms (ICML 2021 · similarity 0.89)
- Blending-Target Domain Adaptation by Adversarial Meta-Adaptation Networks (CVPR 2019 · similarity 0.89)
- A DIRT-T Approach to Unsupervised Domain Adaptation (ICLR 2018 · similarity 0.89)
- Multi-Adversarial Domain Adaptation (AAAI 2018 · 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.