Research map / Label-efficient and robust learning / Domain adaptation and generalization
HyperDomainNet: Universal Domain Adaptation for Generative Adversarial Networks
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 |
|---|---|---|---|
| Replace synthesis-network weight fine-tuning with a 6k-dimensional domain vector | MindTheGap · Face sketches | FID: 78.35 -> 79.83 | No clear effect |
| Replace synthesis-network weight fine-tuning with a 6k-dimensional domain vector | MindTheGap · Face sketches | Precision: 0.326 -> 0.452 | Helped |
| Replace synthesis-network weight fine-tuning with a 6k-dimensional domain vector | MindTheGap · Face sketches | Recall: 0.017 -> 0.017 | No clear effect |
| Add indomain angle consistency loss (ablation) | MindTheGap · Face sketches | FID: 78.35 -> 71.46 | Helped |
| Add indomain angle consistency loss (ablation) | MindTheGap · Face sketches | Precision: 0.326 -> 0.503 | Helped |
| Add indomain angle consistency loss (ablation) | MindTheGap with domain-modulation · Face sketches | Recall: 0.017 -> 0.028 | Helped |
| Use domain-modulation parameterization | StyleGAN-NADA · Anime Painting | Quality: 0.289 -> 0.284 | No clear effect |
| Use domain-modulation parameterization | StyleGAN-NADA · Anime Painting | Diversity: 0.244 -> 0.305 | Helped |
| Add indomain angle consistency loss (ablation) | Ours · Anime Painting | Diversity: 0.305 -> 0.404 | Helped |
| Add indomain angle consistency loss (ablation) | Ours · Anime Painting | Quality: 0.284 -> 0.251 | Hurt |
| Add domain-norm loss (ablation) | HyperDomainNet · Across ten domains | Diversity: 0.099 $\pm$ 0.026 -> 0.306 $\pm$ 0.040 | Helped |
| Add domain-norm loss (ablation) | HyperDomainNet · Across ten domains | Quality: 0.275 $\pm$ 0.035 -> 0.218 $\pm$ 0.026 | Hurt |
Closest papers by meaning
- Adversarial Discriminative Domain Adaptation (CVPR 2017 · similarity 0.91)
- Duplex Generative Adversarial Network for Unsupervised Domain Adaptation (CVPR 2018 · similarity 0.90)
- Unsupervised Pixel-Level Domain Adaptation With Generative Adversarial Networks (CVPR 2017 · similarity 0.90)
- Generate to Adapt: Aligning Domains Using Generative Adversarial Networks (CVPR 2018 · similarity 0.89)
- Mind the Gap: Domain Gap Control for Single Shot Domain Adaptation for Generative Adversarial Networks (ICLR 2022 · similarity 0.89)
- Generalized One-shot Domain Adaptation of Generative Adversarial Networks (NeurIPS 2022 · 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.