Research map / Image generation / GANs
Generalized Loss-Sensitive Adversarial Learning with Manifold Margins
Marzieh Edraki, 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 |
|---|---|---|---|
| replace ambient-only margin with ambient plus manifold margin via pullback mapping (LSAL vs LS-GAN) | LS-GAN · CIFAR-10 generation | Inception score: 5.83 -> 6.43 | Improved |
| LSAL versus ALI for image generation | ALI · CIFAR-10 generation | Inception score: 4.98 -> 6.43 | Improved |
| use LSAL instead of ALI for semi-supervised learning with 1000 labels | ALI · CIFAR-10 semi-supervised (1000 labels) | classification error (%): 19.98 -> 18.83 | Improved |
| LSAL versus ALI with 8000 labels | ALI · CIFAR-10 semi-supervised (8000 labels) | classification error (%): 17.05 -> 14.17 | Improved |
| LSAL versus CLS-GAN with 4000 labels | CLS-GAN · CIFAR-10 semi-supervised (4000 labels) | classification error (%): 17.3 -> 16.22 | Improved |
| LSAL versus CLS-GAN on SVHN with 1000 labels | CLS-GAN · SVHN semi-supervised (1000 labels) | classification error (%): 5.98 -> 5.46 | Improved |
| LSAL versus Improved GAN on SVHN with 1000 labels | Improved GAN · SVHN semi-supervised (1000 labels) | classification error (%): 8.11 -> 5.46 | Improved |
| increase labeled examples from 1000 to 8000 (ablation) | LSAL · CIFAR-10 semi-supervised | classification error (%): 18.83 -> 14.17 | Improved |
Closest papers by meaning
- Manifold Learning Benefits GANs (CVPR 2022 · similarity 0.88)
- Disentangling Adversarial Robustness and Generalization (CVPR 2019 · similarity 0.87)
- Disconnected Manifold Learning for Generative Adversarial Networks (NeurIPS 2018 · similarity 0.87)
- Loss-Sensitive Generative Adversarial Networks on Lipschitz Densities (IJCV 2019 · similarity 0.87)
- SMaRt: Improving GANs with Score Matching Regularity (ICML 2024 · similarity 0.87)
- Enhancing the Adversarial Robustness via Manifold Projection (AAAI 2025 · similarity 0.85)
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.
