Research map / Label-efficient and robust learning / Domain adaptation and generalization
Generalized One-shot Domain Adaptation of 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 |
|---|---|---|---|
| Use the proposed generalized one-shot adaptation framework instead of the strongest baseline | StyleGAN FFHQ adaptation · Sketch, Disney, Arcane, Hat, Zelda | US: 0.13 -> 0.60; 0.11 -> 0.58; 0.11 -> 0.61; 0.10 -> 0.60; 0.11 -> 0.61 | Helped |
| Use the proposed framework instead of the best baseline | StyleGAN FFHQ adaptation · Sketch, Disney, Arcane, Hat, Zelda | NME: 0.09 -> 0.08; 0.11 -> 0.09; 0.08 -> 0.07; 0.08 -> 0.08; 0.09 -> 0.09 | Helped |
| Use the proposed framework instead of the best baseline | StyleGAN FFHQ adaptation · Sketch, Disney, Arcane, Hat, Zelda | IS: 0.31 -> 0.40; 0.19 -> 0.26; 0.24 -> 0.25; 0.29 -> 0.36; 0.13 -> 0.25 | Helped |
| Remove style loss (ablation) | Ours · Sketch | NME: 0.08 -> 0.07 | Helped |
| Remove style loss (ablation) | Ours · Sketch | IS: 0.40 -> 0.45 | Helped |
| Remove style loss (ablation) | Ours · Disney | IS: 0.26 -> 0.40 | Helped |
| Add variational Laplacian regularization (ablation) | Ours without variational Laplacian regularization · Sketch | NME: 0.09 -> 0.08 | Helped |
| Add variational Laplacian regularization (ablation) | Ours without variational Laplacian regularization · Sketch | IS: 0.35 -> 0.40 | Helped |
| Add variational Laplacian regularization (ablation) | Ours without variational Laplacian regularization · Disney | IS: 0.11 -> 0.26 | Helped |
| Remove style loss (ablation) | Ours · adaptation style transfer | style transfer quality | Hurt |
| Remove style fixation (ablation) | Ours · adaptation style transfer | style transfer quality | Hurt |
| Remove entity loss (ablation) | Ours · entity generation | entity generation quality | Hurt |
Closest papers by meaning
- Towards Diverse and Faithful One-shot Adaption of Generative Adversarial Networks (NeurIPS 2022 · similarity 0.95)
- Mind the Gap: Domain Gap Control for Single Shot Domain Adaptation for Generative Adversarial Networks (ICLR 2022 · similarity 0.91)
- Unsupervised Pixel-Level Domain Adaptation With Generative Adversarial Networks (CVPR 2017 · similarity 0.90)
- Adversarial Style Mining for One-Shot Unsupervised Domain Adaptation (NeurIPS 2020 · similarity 0.90)
- HyperDomainNet: Universal Domain Adaptation for Generative Adversarial Networks (NeurIPS 2022 · similarity 0.89)
- Generate to Adapt: Aligning Domains Using Generative Adversarial Networks (CVPR 2018 · 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.