Research map / Image generation / GANs
DO-GAN: A Double Oracle Framework for Generative Adversarial Networks
Check 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 |
|---|---|---|---|
| Add double-oracle training with pruning to vanilla GAN | GAN · CIFAR-10 | Inception Score: +3.36 | Helped |
| Add double-oracle training with pruning to DCGAN | DCGAN · CIFAR-10 | FID: -15.41 | Helped |
| Add double-oracle training with pruning to SNGAN | SNGAN · CIFAR-10 | FID: -7.30 | Helped |
| Add double-oracle training with pruning to SGAN | SGAN · CIFAR-10 | Inception Score: +0.07 | No clear effect |
| Add double-oracle training with continual learning to SGAN | SGAN · CIFAR-10 | Inception Score: +1.16 | Helped |
| Add double-oracle training with pruning to SGAN | SGAN · CelebA | FID: -0.66 | Helped |
| Use continual learning instead of pruning in the DO-SGAN framework (ablation) | DO-SGAN · CIFAR-10 | FID: 16.56 -> 16.07 | Helped |
| Use continual learning instead of pruning in the DO-SGAN framework (ablation) | DO-SGAN · CIFAR-10 | GPU hours: 119.04 -> 97.54 | Helped |
| Set pruning support size to s=5 rather than s=10 (ablation) | DO-GAN/P · 2D Gaussian Dataset | convergence | Hurt |
| Set pruning support size to s=15 rather than s=10 (ablation) | DO-GAN/P · 2D Gaussian Dataset | runtime (GPU hours): 0.5627 -> 0.9989 | Hurt |
| Use DO-GAN/P rather than vanilla GAN | GAN · 2D Gaussian mixture | number of modes generated: 6 out of 8 -> all the 8 | Helped |
Closest papers by meaning
- Finding Mixed Nash Equilibria of Generative Adversarial Networks (ICML 2019 · similarity 0.92)
- An Online Learning Approach to Generative Adversarial Networks (ICLR 2018 · similarity 0.92)
- The Numerics of GANs (NeurIPS 2017 · similarity 0.92)
- Generative Multi-Adversarial Networks (ICLR 2017 · similarity 0.91)
- Are GANs Created Equal? A Large-Scale Study (NeurIPS 2018 · similarity 0.91)
- Training Generative Adversarial Networks with Limited Data (NeurIPS 2020 · similarity 0.91)
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.