Research map / Label-efficient and robust learning / Recognition and long-tail
Teaching a GAN What Not to Learn
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 |
|---|---|---|---|
| Train with positive and negative samples, and teach the generator to avoid the negative class | LSGAN · MNIST unbalanced classes (averaged) | FID: 121.3 -> 118.93 | Helped |
| Train with positive and negative samples, and teach the generator to avoid the negative class | LSGAN · CelebA unbalanced classes (averaged) | FID: 213.6 -> 169.34 | Helped |
| Train with positive and negative samples, and teach the generator to avoid the negative class | LSGAN · CIFAR-10 unbalanced classes (averaged) | FID: 231.02 -> 217.15 | Helped |
| Train on positive and negative classes while directing generation toward the positive class | LSGAN · MNIST even digits and overlapping classes | FID | Helped |
| Train on positive and negative classes while directing generation toward the positive class | LSGAN · Fashion-MNIST overlapping classes | precision-recall | Helped |
| Use the Rumi formulation rather than the standard GAN objective | SGAN · MNIST | FID | Helped |
| Use the Rumi formulation rather than the standard LSGAN objective | LSGAN · MNIST | FID | Helped |
| Include negative-class samples and train the generator to avoid that class | LSGAN · MNIST minority-class learning | precision-recall | Helped |
| Include negative-class samples and train the generator to avoid that class | LSGAN · CIFAR-10 animals versus vehicles | FID | Helped |
Closest papers by meaning
- Training Generative Adversarial Networks with Limited Data (NeurIPS 2020 · similarity 0.90)
- SGAN: An Alternative Training of Generative Adversarial Networks (CVPR 2018 · similarity 0.89)
- GLeaD: Improving GANs With a Generator-Leading Task (CVPR 2023 · similarity 0.89)
- Sequential Training of GANs Against GAN-Classifiers Reveals Correlated "Knowledge Gaps" Present Among Independently Trained GAN Instances (CVPR 2023 · similarity 0.89)
- Taming the Tail in Class-Conditional GANs: Knowledge Sharing via Unconditional Training at Lower Resolutions (CVPR 2024 · similarity 0.89)
- SAN: Inducing Metrizability of GAN with Discriminative Normalized Linear Layer (ICLR 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.