Research map / Optimization / Optimizers
Sinkhorn Natural Gradient for Generative Models
Zebang Shen, Zhenfu Wang, Alejandro Ribeiro, Hamed Hassani
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (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 SiNG instead of Adam | DC-GAN generator · CelebA | generator loss | Improved |
| Use SiNG instead of SGD-type optimizers | DC-GAN generator with a fixed encoder ground cost · CelebA | generator loss | Improved |
| Use SiNG instead of SGD-type optimizers | DC-GAN generator with a fixed encoder ground cost · CIFAR-10 | generator loss | Improved |
| Use SiNG instead of SGD-based optimizers | DC-GAN · GAN image generation | visual quality | Improved |
| Use the kernelized Wasserstein natural gradient (KWNG) | generative-model optimization · paper's experimental setting | convergence | Hurt |
Closest papers by meaning
- Divergence Frontiers for Generative Models: Sample Complexity, Quantization Effects, and Frontier Integrals (NeurIPS 2021 · similarity 0.83)
- An Architecture for Deep, Hierarchical Generative Models (NeurIPS 2016 · similarity 0.82)
- Auxiliary Deep Generative Models (ICML 2016 · similarity 0.82)
- Don’t Generate Me: Training Differentially Private Generative Models with Sinkhorn Divergence (NeurIPS 2021 · similarity 0.82)
- Sinkhorn Barycenter via Functional Gradient Descent (NeurIPS 2020 · similarity 0.82)
- Statistically Optimal Generative Modeling with Maximum Deviation from the Empirical Distribution (ICML 2024 · similarity 0.80)
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.
