Research map / Image generation / GANs
Loss-Sensitive Generative Adversarial Networks on Lipschitz Densities
Guo-Jun Qi
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (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 CLS-GAN loss-sensitive objective instead of DCGAN features | CLS-GAN · CIFAR-10 (all labels) | accuracy: 82.8% -> 91.7% | Improved |
| use CLS-GAN instead of Improved GAN with 400 labels per class | CLS-GAN · CIFAR-10 (400 labels per class) | accuracy: 81.4% -> 82.7% | Improved |
| use CLS-GAN instead of ALI on SVHN | CLS-GAN · SVHN (1000 labels) | error rate: 7.3% -> 5.98% | Improved |
| remove batch normalization from generator (ablation) | DCGAN · CelebA generation | image quality | Hurt |
| remove batch normalization from generator (ablation) | LS-GAN · CelebA generation | image quality | No clear effect |
| update generator every 3 or 5 iterations while loss updated every iteration (ablation) | LS-GAN · CelebA generation | image quality | No clear effect |
| use L1 distance of image features as margin; use external Inception/VGG-16 features (ablation) | LS-GAN · CelebA generation | image quality | Improved |
| halve lambda (ablation) | CLS-GAN · MNIST, CIFAR-10, SVHN generation | sample diversity | Hurt |
| use regularized GAN (WGAN) instead of DCGAN | WGAN · CIFAR-10 test MRE | MRE: 0.1506 -> 0.1109 | Improved |
| use GLS-GAN instead of DCGAN | GLS-GAN · CIFAR-10 test MRE | MRE: 0.1506 -> 0.1089 | Improved |
| use GLS-GAN with LeakyReLU slope 0.5 on tiny ImageNet | GLS-GAN · tiny ImageNet test MRE | MRE: 0.2085 | Improved |
Closest papers by meaning
- Lipschitz Generative Adversarial Nets (ICML 2019 · similarity 0.92)
- Adversarial Lipschitz Regularization (ICLR 2020 · similarity 0.91)
- How does Lipschitz Regularization Influence GAN Training? (ECCV 2020 · similarity 0.91)
- On the regularization of Wasserstein GANs (ICLR 2018 · similarity 0.91)
- Regularizing Generative Adversarial Networks Under Limited Data (CVPR 2021 · similarity 0.91)
- MCGAN: Enhancing GAN Training with Regression-Based Generator Loss (AAAI 2025 · 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.
