Research map / Label-efficient and robust learning / Recognition and long-tail
Conditional GANs with Auxiliary Discriminative Classifier
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 a generator-aware auxiliary discriminative classifier instead of the standard classifier | Conditional GAN · CIFAR-10 | FID: 5.83 -> 5.66 | Helped |
| Use a generator-aware auxiliary discriminative classifier instead of the standard classifier | Conditional GAN · CIFAR-100 | Intra-FID: 51.15 -> 49.24 | Helped |
| Use a generator-aware auxiliary discriminative classifier instead of the standard classifier | Conditional GAN · Tiny-ImageNet | FID: 21.12 -> 19.02 | Helped |
| Use a generator-aware auxiliary discriminative classifier instead of the standard classifier | Conditional GAN · ImageNet 128×128, batch size 256, 500k iterations | IS: 68.27 -> 66.96 | Hurt |
| Use a generator-aware auxiliary discriminative classifier instead of the standard classifier | Conditional GAN · ImageNet 128×128, batch size 256, 500k iterations | FID: 13.98 -> 11.65 | Helped |
| Use a generator-aware auxiliary discriminative classifier instead of the standard classifier | Conditional GAN · ImageNet 128×128, batch size 2048, 200k iterations | FID: 7.89 -> 9.46 | Hurt |
| Remove the original GAN loss (ablation) | ADC-GAN · One-dimensional synthetic mixture of Gaussians | joint distribution recovery | No clear effect |
| Remove the original discriminator (ablation) | ADC-GAN · CIFAR-10, CIFAR-100, and Tiny-ImageNet | training stability | No clear effect |
| Increase the classifier-task coefficient (ablation) | ADC-GAN · CIFAR-100 | FID | No clear effect |
| Remove the discriminator from AC-GAN on synthetic data (ablation) | AC-GAN · One-dimensional synthetic mixture of Gaussians | intra-class diversity | Hurt |
| Remove the GAN loss from TAC-GAN (ablation) | TAC-GAN · One-dimensional synthetic mixture of Gaussians | distribution recovery | Hurt |
Closest papers by meaning
- Twin Auxilary Classifiers GAN (NeurIPS 2019 · similarity 0.94)
- Rebooting ACGAN: Auxiliary Classifier GANs with Stable Training (NeurIPS 2021 · similarity 0.90)
- Label-Noise Robust Generative Adversarial Networks (CVPR 2019 · similarity 0.89)
- Taming the Tail in Class-Conditional GANs: Knowledge Sharing via Unconditional Training at Lower Resolutions (CVPR 2024 · similarity 0.88)
- Discriminative Forests Improve Generative Diversity for Generative Adversarial Networks (AAAI 2024 · similarity 0.88)
- Fusing Conditional Submodular GAN and Programmatic Weak Supervision (AAAI 2024 · similarity 0.87)
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.