Research map / Image generation / GANs
Robust Conditional GAN from Uncertainty-Aware Pairwise Comparisons
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 pairwise-comparison supervision instead of full supervision | PC-GAN · CACD | classification accuracy (%): 48.44 vs 46.02 | Mixed |
| Use pairwise-comparison supervision instead of full supervision | PC-GAN · UTKFace | classification accuracy (%): 63.88 vs 71.44 | Hurt |
| Use pairwise-comparison supervision instead of full supervision | PC-GAN · SCUT-FBP | classification accuracy (%): 40.00 vs 46.25 | Hurt |
| Add rating reconstruction loss (ablation) | PC-GAN · CACD | classification accuracy (%): 39.50 -> 48.08 | Helped |
| Add rating reconstruction loss (ablation) | PC-GAN · UTKFace | classification accuracy (%): 56.90 -> 62.74 | Helped |
| Remove cycle loss (ablation) | PC-GAN · CACD | classification accuracy (%): 48.08 -> 50.86 | Helped |
| Remove cycle loss (ablation) | PC-GAN · UTKFace | classification accuracy (%): 62.74 -> 60.56 | Hurt |
| Remove identity-preserving loss (ablation) | PC-GAN · CACD | classification accuracy (%): 48.08 -> 48.60 | No clear effect |
| Remove identity-preserving loss (ablation) | PC-GAN · UTKFace | classification accuracy (%): 62.74 -> 63.92 | No clear effect |
| Use Bayesian uncertainty estimation and noise-robust conditional training (ablation) | PC-GAN · CACD | classification accuracy (%): 35.04 -> 41.02 | Helped |
| Use Bayesian uncertainty estimation and noise-robust conditional training (ablation) | PC-GAN · UTKFace | classification accuracy (%): 40.12 -> 43.64 | Helped |
| Filter equal-intensity pairs from CGAN training (ablation) | PC-GAN · UTKFace | classification accuracy (%): 58.10 -> 67.44 | Helped |
Closest papers by meaning
- Robust Conditional Generative Adversarial Networks (ICLR 2019 · similarity 0.89)
- Robustness of conditional GANs to noisy labels (NeurIPS 2018 · similarity 0.89)
- Label-Noise Robust Generative Adversarial Networks (CVPR 2019 · similarity 0.89)
- Generative Attribute Controller With Conditional Filtered Generative Adversarial Networks (CVPR 2017 · similarity 0.88)
- UQGAN: A Unified Model for Uncertainty Quantification of Deep Classifiers trained via Conditional GANs (NeurIPS 2022 · similarity 0.88)
- Taming the Tail in Class-Conditional GANs: Knowledge Sharing via Unconditional Training at Lower Resolutions (CVPR 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.