Research map / Label-efficient and robust learning / Self-supervised and contrastive
Self-supervised GAN: Analysis and Improvement with Multi-class Minimax Game
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 |
|---|---|---|---|
| Replace original SS tasks with multi-class minimax SS tasks | Dist-GAN, SN-GAN architecture · CIFAR-10 | FID: 21.40 -> 18.88 | Helped |
| Replace original SS tasks with multi-class minimax SS tasks | Dist-GAN, SN-GAN architecture · STL-10 | FID: 29.79 -> 27.95 | Helped |
| Replace original SS tasks with multi-class minimax SS tasks | Dist-GAN, ResNet · CIFAR-10 | FID: 14.97 ± .29 -> 13.90 ± .22 | Helped |
| Replace original SS tasks with multi-class minimax SS tasks | Dist-GAN, ResNet · STL-10 | FID: 27.98 ± .38 -> 27.10 ± .34 | Helped |
| Replace original SS tasks with multi-class minimax SS tasks | Dist-GAN, ResNet · CIFAR-10, 10K-10K FID | FID: 12.37 -> 11.40 | Helped |
| Replace original SS tasks with multi-class minimax SS tasks | Dist-GAN · CIFAR-100 | FID: 21.02 -> 19.74 | Helped |
| Replace original SS tasks with multi-class minimax SS tasks | Dist-GAN · ImageNet 32×32 | FID: 17.1 -> 12.3 | Helped |
| Replace original SS tasks with multi-class minimax SS tasks | Dist-GAN · Stacked-MNIST, K/4 | covered modes: 906.75 ± 26.15 -> 926.75 ± 32.65 | Helped |
| Replace original SS tasks with multi-class minimax SS tasks | Dist-GAN · Stacked-MNIST, K/4 | KL divergence: 0.90 ± 0.13 -> 0.78 ± 0.13 | Helped |
| Replace original SS tasks with multi-class minimax SS tasks | Dist-GAN · Stacked-MNIST, K/2 | covered modes: 957.50 ± 31.23 -> 976.00 ± 10.04 | Helped |
| Replace original SS tasks with multi-class minimax SS tasks | Dist-GAN · Stacked-MNIST, K/2 | KL divergence: 0.61 ± 0.15 -> 0.52 ± 0.07 | Helped |
| Replace original SS tasks with multi-class minimax SS tasks | SSGAN+ · CelebA 64×64 | FID: 35.03 -> 33.47 | Helped |
Closest papers by meaning
- Self-Supervised GANs with Label Augmentation (NeurIPS 2021 · similarity 0.92)
- Self-Supervised GANs via Auxiliary Rotation Loss (CVPR 2019 · similarity 0.91)
- A Unified Generative Adversarial Network Training via Self-Labeling and Self-Attention (ICML 2021 · similarity 0.90)
- On Self-Supervised Image Representations for GAN Evaluation (ICLR 2021 · similarity 0.90)
- MCGAN: Enhancing GAN Training with Regression-Based Generator Loss (AAAI 2025 · similarity 0.88)
- Taming the Tail in Class-Conditional GANs: Knowledge Sharing via Unconditional Training at Lower Resolutions (CVPR 2024 · 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.