Research map / Learning theory and algorithms / Sample complexity
Minimax Optimality (Probably) Doesn't Imply Distribution Learning for GANs
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 a local-PRG-based generator with a Gaussian seed instead of sampling the target distribution | ReLU GAN generator · Diverse continuous target distributions | Wasserstein-1 distance: ≥ Ω(1) | Hurt |
| Use a local-PRG-based generator with a Gaussian seed instead of sampling the target distribution | ReLU GAN generator · Diverse continuous target distributions | Polynomial-size ReLU discriminator IPM: ≤ ε(m)·poly(m) | No clear effect |
| Generate uniform-hypercube samples using a local PRG | Local-PRG generator G_m · Uniform distribution U_d | Wasserstein-1 distance: ≥ 1 | Hurt |
| Generate uniform-hypercube samples using a local PRG | Local-PRG generator G_m · Uniform distribution U_d | Polynomial-size ReLU discriminator IPM: ≤ negl(m) | No clear effect |
| Construct an explicit generator that fools the specified ReLU discriminator class | Continuous-domain GAN generator · Uniform distribution I_d | ReLU discriminator IPM: ≤ ε(m) | Mixed |
Closest papers by meaning
- Approximation and Convergence Properties of Generative Adversarial Learning (NeurIPS 2017 · similarity 0.89)
- On the Convergence and Robustness of Training GANs with Regularized Optimal Transport (NeurIPS 2018 · similarity 0.89)
- Do GANs learn the distribution? Some Theory and Empirics (ICLR 2018 · similarity 0.88)
- Hidden Convexity of Wasserstein GANs: Interpretable Generative Models with Closed-Form Solutions (ICLR 2022 · similarity 0.88)
- Statistically Optimal Generative Modeling with Maximum Deviation from the Empirical Distribution (ICML 2024 · similarity 0.87)
- SAN: Inducing Metrizability of GAN with Discriminative Normalized Linear Layer (ICLR 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.