Research map / Label-efficient and robust learning / Domain adaptation and generalization
Conditional Coupled Generative Adversarial Networks for Zero-Shot Domain Adaptation
Jinghua Wang, Jianmin Jiang
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 RT/IRT task classifiers for representation alignment (ablation) | CoCoGAN w/o T · Zero-shot DA on MNIST/Fashion-MNIST/NIST/EMNIST, average | accuracy (%): +15.6% on average | Improved |
| add task classifiers (NIST RT, IRT Fashion-MNIST, color->gray) (ablation) | CoCoGAN w/o T · NIST, color->gray | accuracy (%): 30.9 -> 57.7 | Improved |
| add task classifiers (MNIST RT, IRT EMNIST, color->gray) (ablation) | CoCoGAN w/o T · MNIST, color->gray | accuracy (%): 63.5 -> 94.7 | Improved |
| add task classifiers (MNIST RT, IRT Fashion-MNIST, gray->color) (ablation) | CoCoGAN w/o T · MNIST, gray->color | accuracy (%): 68.3 -> 78.1 | Improved |
| add task classifiers (Fashion-MNIST RT, IRT NIST, gray->color) (ablation) | CoCoGAN w/o T · Fashion-MNIST, gray->color | accuracy (%): 39.7 -> 56.8 | Mixed |
| add task classifiers (NIST RT, IRT MNIST, gray->negative) (ablation) | CoCoGAN w/o T · NIST, gray->negative | accuracy (%): 39.3 -> 45.7 | Improved |
| add task classifiers to improve generated-image correspondence (ablation) | CoCoGAN w/o T · MNIST, gray->edge | overlap ratio: 0.816 -> 0.873 | Improved |
| add task classifiers to improve generated-image correspondence (ablation) | CoCoGAN w/o T · MNIST, gray->negative | overlap ratio: 0.804 -> 0.863 | Improved |
| use CoCoGAN instead of ZDDA, NIST with digit IRT (color->gray) | ZDDA · NIST, color->gray, IRT MNIST | accuracy (%): +7.9% | Improved |
| use CoCoGAN instead of ZDDA, NIST with fashion IRT (color->gray) | ZDDA · NIST, color->gray, IRT Fashion-MNIST | accuracy (%): 23.7 -> 57.7 | Improved |
| use CoCoGAN instead of ZDDA, MNIST with EMNIST IRT (gray->color) | ZDDA · MNIST, gray->color | accuracy (%): 94.8 -> 95.6 | Improved |
| use CoCoGAN with no target data versus methods that use target data (color->gray MNIST) | existing DA methods with target data · MNIST, color->gray | accuracy (%): 94.2 -> 94.7 | Improved |
Closest papers by meaning
- Adversarial Learning for Zero-shot Domain Adaptation (ECCV 2020 · similarity 0.92)
- Learning Across Tasks for Zero-Shot Domain Adaptation From a Single Source Domain (TPAMI 2021 · similarity 0.92)
- Coupled Generative Adversarial Networks (NeurIPS 2016 · similarity 0.90)
- Transferring GANs: generating images from limited data (ECCV 2018 · similarity 0.90)
- One-Shot Generative Domain Adaptation (ICCV 2023 · similarity 0.89)
- Turning Waste into Wealth: Leveraging Low-Quality Samples for Enhancing Continuous Conditional Generative Adversarial Networks (AAAI 2024 · similarity 0.88)
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.
