Research map / Adversarial robustness and security / Certified and empirical robustness
Regularizing Generative Adversarial Networks Under Limited Data
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 LeCam regularization to BigGAN | BigGAN · CIFAR-10, 20% data | FID: 21.86±0.29 -> 15.27±0.10 | Helped |
| Add LeCam regularization to BigGAN | BigGAN · CIFAR-100, 20% data | FID: 32.99±0.24 -> 25.51±0.19 | Helped |
| Add LeCam regularization to BigGAN | BigGAN · ImageNet, 25% data | FID: 18.22±2.59 -> 14.47±1.73 | Helped |
| Add LeCam regularization to DA | BigGAN + DA · ImageNet, 10% data | FID: 37.71±0.11 -> 24.38±0.06 | Helped |
| Add LeCam regularization to StyleGAN2 | StyleGAN2 · StyleGAN dataset, 1K images | FID: 72.07±0.04 -> 63.16±0.11 | Helped |
| Regularize both real-image and generated-image predictions instead of only real-image predictions (ablation) | BigGAN · CIFAR-10, 20% data | FID: 20.47±0.36 -> 15.27±0.10 | Helped |
| Regularize both real-image and generated-image predictions instead of only generated-image predictions (ablation) | BigGAN · CIFAR-10, 20% data | FID: 18.18±0.08 -> 15.27±0.10 | Helped |
| Replace moving-average anchors with fixed anchors 0.5 and -0.5 (ablation) | BigGAN + LeCam regularization · CIFAR-10, 20% data | FID: 15.27±0.10 -> 30.64±0.05 | Hurt |
| Replace moving-average anchors with fixed anchors 1 and -1 (ablation) | BigGAN + LeCam regularization · CIFAR-10, 20% data | FID: 15.27±0.10 -> 19.81±0.03 | Hurt |
| Increase regularization weight above 0.5 (ablation) | BigGAN + LeCam regularization · CIFAR-10, 20% data | FID | Hurt |
Closest papers by meaning
- Stabilizing Training of Generative Adversarial Networks through Regularization (NeurIPS 2017 · similarity 0.93)
- A Large-Scale Study on Regularization and Normalization in GANs (ICML 2019 · similarity 0.91)
- Training Generative Adversarial Networks with Limited Data (NeurIPS 2020 · similarity 0.91)
- MS$^3$D: A RG Flow-Based Regularization for GAN Training with Limited Data (ICML 2024 · similarity 0.90)
- Consistency Regularization for Generative Adversarial Networks (ICLR 2020 · similarity 0.90)
- Mode Regularized Generative Adversarial Networks (ICLR 2017 · similarity 0.90)
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.