Research map / Image generation / GANs
Disconnected Manifold Learning 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 |
|---|---|---|---|
| Learn the prior over generators instead of fixing it | DMWGAN · MNIST | JSD ×10^-2: 0.23 std 0.06 -> 0.06 std 0.02 | Helped |
| Learn the prior over generators instead of fixing it | DMWGAN · Face-Bed | JSD ×10^-4: 0.46 std 0.25 -> 0.10 std 0.05 | Helped |
| Learn the prior over generators instead of fixing it | DMWGAN · Face-Bed | FID: 7.96 std 0.08 -> 7.67 std 0.16 | Helped |
| Add the mutual information term (ablation) | DMWGAN-PL · MNIST | JSD ×10^-2: 0.13 std 0.01 -> 0.06 std 0.02 | Helped |
| Add the mutual information term (ablation) | DMWGAN-PL · Face-Bed | JSD ×10^-4: 0.26 std 0.11 -> 0.10 std 0.05 | Helped |
| Add the mutual information term (ablation) | DMWGAN-PL · Face-Bed | FID: 7.80 std 0.14 -> 7.67 std 0.16 | No clear effect |
| Use multiple generators with learned prior rather than a single generator | GAN-GP · MNIST, modified GAN objective | JSD: 0.0052 std 0.0007 -> 0.0005 std 0.0002 | Helped |
Closest papers by meaning
- Learning disconnected manifolds: a no GAN’s land (ICML 2020 · similarity 0.95)
- Manifold Learning Benefits GANs (CVPR 2022 · similarity 0.89)
- A Geometric Analysis of Deep Generative Image Models and Its Applications (ICLR 2021 · similarity 0.89)
- Global Versus Localized Generative Adversarial Nets (CVPR 2018 · similarity 0.89)
- Learning Generative Models across Incomparable Spaces (ICML 2019 · similarity 0.88)
- SMaRt: Improving GANs with Score Matching Regularity (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.