Research map / Label-efficient and robust learning / Domain adaptation and generalization
From Source to Target and Back: Symmetric Bi-Directional Adaptive GAN
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 |
|---|---|---|---|
| Combine the two classifier outputs | SBADA-GAN · MNIST → USPS | accuracy: 97.1 -> 97.6 | Helped |
| Combine the two classifier outputs | SBADA-GAN · USPS → MNIST | accuracy: 87.5 -> 95.0 | Helped |
| Combine the two classifier outputs | SBADA-GAN · MNIST → MNIST-M | accuracy: 98.4 -> 99.4 | Helped |
| Combine the two classifier outputs | SBADA-GAN · SVHN → MNIST | accuracy: 74.2 -> 76.1 | Helped |
| Combine the two classifier outputs | SBADA-GAN · MNIST → SVHN | accuracy: 59.2 -> 61.1 | Helped |
| Combine the two classifier outputs | SBADA-GAN · Synth Signs → GTSRB | accuracy: 95.9 -> 96.7 | Helped |
| Add a target-to-source GAN alongside the source-to-target GAN (ablation) | SBADA-GAN · MNIST → USPS | accuracy: 94.23 -> 94.90 | Helped |
| Add class consistency between the two generators (ablation) | SBADA-GAN · MNIST → USPS | accuracy: 94.90 -> 95.45 | Helped |
| Add target self-labeling (ablation) | SBADA-GAN · MNIST → USPS | accuracy: 95.45 -> 97.60 | Helped |
| Add target self-labeling (ablation) | SBADA-GAN · MNIST ↔ SVHN | accuracy | No clear effect |
| Use both mapping directions and their classifiers | SBADA-GAN · SVHN → MNIST | accuracy: 76.1 vs 97.6 | Hurt |
| Use both mapping directions and their classifiers | SBADA-GAN · MNIST → SVHN | accuracy: 61.1 vs 52.8 | Helped |
Closest papers by meaning
- Unsupervised Pixel-Level Domain Adaptation With Generative Adversarial Networks (CVPR 2017 · similarity 0.90)
- Self-Supervised GANs via Auxiliary Rotation Loss (CVPR 2019 · similarity 0.88)
- AdaptiveMix: Improving GAN Training via Feature Space Shrinkage (CVPR 2023 · similarity 0.88)
- Instance-Conditioned GAN (NeurIPS 2021 · similarity 0.88)
- Transformation GAN for Unsupervised Image Synthesis and Representation Learning (CVPR 2020 · similarity 0.87)
- E$^2$GAN: Efficient Training of Efficient GANs for Image-to-Image Translation (ICML 2024 · similarity 0.86)
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.