Research map / Label-efficient and robust learning / Domain adaptation and generalization
Dual Mixup Regularized Learning for Adversarial Domain Adaptation
Yuan Wu, Diana Inkpen, Ahmed El-Roby
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 domain mixup (ablation) | DMRL · Digits MNIST to USPS | accuracy (%): 94.8 -> 96.1 | Improved |
| add domain mixup (ablation) | DMRL · Digits USPS to MNIST | accuracy (%): 97.8 -> 99.0 | Improved |
| add category mixup (ablation) | DMRL · Digits MNIST to USPS | accuracy (%): 90.3 -> 96.1 | Improved |
| add category mixup (ablation) | DMRL · Digits USPS to MNIST | accuracy (%): 92.6 -> 99.0 | Improved |
| add labeled category mixup (ablation) | DMRL · Digits MNIST to USPS | accuracy (%): 95.0 -> 96.1 | Improved |
| add labeled category mixup (ablation) | DMRL · Digits USPS to MNIST | accuracy (%): 97.9 -> 99.0 | Improved |
| add unlabeled category mixup (ablation) | DMRL · Digits MNIST to USPS | accuracy (%): 90.7 -> 96.1 | Improved |
| add unlabeled category mixup (ablation) | DMRL · Digits USPS to MNIST | accuracy (%): 93.3 -> 99.0 | Improved |
| add domain adaptation (full DMRL vs no adaptation) | DMRL · Digits MNIST to USPS | accuracy (%): 82.2 -> 96.1 | Improved |
| add domain adaptation (full DMRL vs source-only ResNet-50) | DMRL · Office-31 (average of 6 tasks) | accuracy (%): 76.1 -> 87.9 | Improved |
| compare DMRL with CDAN on balanced domains | DMRL · ImageCLEF-DA (average of 6 tasks) | accuracy (%): 87.1 -> 88.0 | Improved |
| increase lambda_t from 0.1 up to 2.0 then beyond (ablation) | DMRL · Digits MNIST to USPS | accuracy (%): best at lambda_t=2.0 | Mixed |
Closest papers by meaning
- Adversarial Domain Adaptation with Domain Mixup (AAAI 2020 · similarity 0.93)
- On Unsupervised Domain Adaptation: Pseudo Label Guided Mixup for Adversarial Prompt Tuning (AAAI 2024 · similarity 0.90)
- Connect, Not Collapse: Explaining Contrastive Learning for Unsupervised Domain Adaptation (ICML 2022 · similarity 0.89)
- Cross-Domain Gradient Discrepancy Minimization for Unsupervised Domain Adaptation (CVPR 2021 · similarity 0.88)
- Divergence-Agnostic Unsupervised Domain Adaptation by Adversarial Attacks (TPAMI 2021 · similarity 0.88)
- Can We Evaluate Domain Adaptation Models Without Target-Domain Labels? (ICLR 2024 · 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.
