Research map / Label-efficient and robust learning / Domain adaptation and generalization
Mind the Discriminability: Asymmetric Adversarial Domain Adaptation
Jianfei Yang, Han Zou, Yuxun Zhou, Zhaoyang Zeng, Lihua Xie ()
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 |
|---|---|---|---|
| replace symmetric domain classifier with asymmetric autoencoder discriminator (AADA vs DANN) | DANN · SVHN->MNIST | accuracy (%): 74.7 -> 98.1 | Improved |
| replace symmetric domain classifier with asymmetric autoencoder discriminator (AADA vs DANN) | DANN · Office-Home | accuracy (%): 57.6 -> 65.3 | Improved |
| replace symmetric domain classifier with asymmetric autoencoder discriminator (AADA vs DANN) | DANN · Image-CLEF | accuracy (%): 85.0 -> 86.5 | Improved |
| integrate AADA with CCN | CCN · Image-CLEF | accuracy (%): 85.1 -> 88.4 | Improved |
| integrate AADA with CCN | CCN · Office-Home | accuracy (%): 59.7 -> 67.0 | Improved |
| fix source and push target (AADA) rather than fix target and push source (AADAopt) (ablation) | AADA · SVHN->MNIST | accuracy (%): 74.8 -> 98.1 | Improved |
| fix source and push target (AADA) rather than fix target and push source (AADAopt) (ablation) | AADA · MNIST->MNIST-M | accuracy (%): 47.1 -> 95.5 | Improved |
| fix source and push target (AADA) rather than fix target and push source (AADAopt) (ablation) | AADA · MNIST->USPS | accuracy (%): 95.6 -> 98.4 | Improved |
| increase margin m (ablation) | AADA · SVHN->MNIST | accuracy (%) | Improved |
| use large adversarial weight gamma (ablation) | AADA vs DANN · SVHN->MNIST | accuracy (%) | Mixed |
| use AADA instead of DANN for spectral discriminability | AADA · SVHN->MNIST | ideal joint hypothesis lambda | Improved |
Closest papers by meaning
- Adversarial Discriminative Domain Adaptation (CVPR 2017 · similarity 0.92)
- Reusing the Task-Specific Classifier as a Discriminator: Discriminator-Free Adversarial Domain Adaptation (CVPR 2022 · similarity 0.91)
- Adversarial-Learned Loss for Domain Adaptation (AAAI 2020 · similarity 0.91)
- Domain-Adversarial Training of Neural Networks (JMLR 2016 · similarity 0.90)
- Adversarial Domain Adaptation with Domain Mixup (AAAI 2020 · similarity 0.90)
- Bridging Theory and Algorithm for Domain Adaptation (ICML 2019 · 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.
