Research map / Image generation / GANs
Stein Latent Optimization 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 |
|---|---|---|---|
| Use SLOGAN rather than the strongest listed baseline | unsupervised conditional GAN · CIFAR-2 (7:3) | NMI: 0.42 -> 0.69 | Helped |
| Use SLOGAN rather than the strongest listed baseline | unsupervised conditional GAN · CIFAR-2 (7:3) | ICFID: 57.06 -> 45.83 | Helped |
| Use SLOGAN rather than the strongest listed baseline | unsupervised conditional GAN · CIFAR-10 | FID: 31.50 -> 20.61 | Helped |
| Add U2C loss (ablation) | SLOGAN · MNIST-2 (7:3) | ICFID: 9.43 -> 5.91 | Helped |
| Add U2C loss (ablation) | SLOGAN · CIFAR-2 (7:3) | NMI: 0.08 -> 0.69 | Helped |
| Update μ, Σ, and ρ and use U2C loss (ablation) | SLOGAN · CIFAR-2 (7:3) | ICFID: 84.44 -> 45.83 | Helped |
| Retain μ update (ablation) | SLOGAN · CIFAR-2 (7:3) | ICFID: 73.79 -> 45.83 | Helped |
| Retain ρ update (ablation) | SLOGAN · CIFAR-2 (7:3) | ICFID: 63.09 -> 45.83 | Helped |
| Use implicit rather than explicit reparameterization for the mixture parameters (ablation) | SLOGAN · CIFAR-2 (7:3) | ICFID: 60.51 -> 45.83 | Helped |
| Add implicit ρ update to the DeLiGAN variant (ablation) | DeLiGAN with U2C loss · CIFAR-2 (7:3) | ICFID: 60.51 -> 86.48 | Hurt |
| Do not use U2C loss to update ρ (ablation) | SLOGAN · CIFAR-2 (7:3) | ICFID: 52.67 -> 45.83 | Helped |
| Add SimCLR loss (ablation) | SLOGAN · CIFAR-2 (7:3) | ICFID: 49.25 -> 45.83 | Helped |
Closest papers by meaning
- InfoGAN-CR and ModelCentrality: Self-supervised Model Training and Selection for Disentangling GANs (ICML 2020 · similarity 0.89)
- AdaGAN: Boosting Generative Models (NeurIPS 2017 · similarity 0.89)
- Variational Annealing of GANs: A Langevin Perspective (ICML 2019 · similarity 0.89)
- ClusterGAN: Latent Space Clustering in Generative Adversarial Networks (AAAI 2019 · similarity 0.89)
- Taming the Tail in Class-Conditional GANs: Knowledge Sharing via Unconditional Training at Lower Resolutions (CVPR 2024 · similarity 0.88)
- MCGAN: Enhancing GAN Training with Regression-Based Generator Loss (AAAI 2025 · 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.