Research map / Image generation / GANs
AdaptiveMix: Improving GAN Training via Feature Space Shrinkage
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 |
|---|---|---|---|
| Add AdaptiveMix to WGAN training | WGAN · CIFAR-10 | Min FID: 55.96 -> 30.85 | Helped |
| Add AdaptiveMix to StyleGAN-V2 | StyleGAN-V2 · AFHQ-Cat-5k | FID: 7.737 -> 4.477 | Helped |
| Add AdaptiveMix to APA | APA with StyleGAN-V2 · AFHQ-Cat-5k | FID: 4.876 -> 4.148 | Helped |
| Add AdaptiveMix to StyleGAN-V2 | StyleGAN-V2 · FFHQ-5k | FID: 37.830 -> 18.769 | Helped |
| Add AdaptiveMix to APA | APA with StyleGAN-V2 · FFHQ-5k | FID: 13.249 -> 11.498 | Helped |
| Add AdaptiveMix to DCGAN's WGAN objective | DCGAN · CelebA | FID: 23.51 -> 12.43 | Helped |
| Add AdaptiveMix to image classification | WRN-28-10 · CIFAR-10 | accuracy: 96.11 -> 96.80 | Helped |
| Add AdaptiveMix to image classification | PreActResNet-18 · Tiny-ImageNet | accuracy: 57.23 -> 60.59 | Helped |
| Add AdaptiveMix to 1DS without Monte Carlo sampling | 1DS w/o MC · CIFAR-10 / TIN-C | F1 score: 0.890 -> 0.922 | Helped |
| Set the mixing Beta-distribution parameter α to 2.0 instead of 1.0 (ablation) | AdaptiveMix with WRN-28-10 · CIFAR-10 | accuracy (FGSM 8/255): 74.18 -> 71.27 | Hurt |
| Add AdaptiveMix to APA with StyleGAN-V2 | APA · FFHQ (Full) | FID: 3.678 -> 3.609 | Helped |
Closest papers by meaning
- Re-GAN: Data-Efficient GANs Training via Architectural Reconfiguration (CVPR 2023 · similarity 0.90)
- Improved Techniques for Training GANs (NeurIPS 2016 · similarity 0.89)
- MCGAN: Enhancing GAN Training with Regression-Based Generator Loss (AAAI 2025 · similarity 0.89)
- RG-GAN: Dynamic Regenerative Pruning for Data-Efficient Generative Adversarial Networks (AAAI 2024 · similarity 0.88)
- Direct Discriminative Optimization: Your Likelihood-Based Visual Generative Model is Secretly a GAN Discriminator (ICML 2025 · similarity 0.88)
- Revisiting Adversarial Training at Scale (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.