Research map / Image generation / GANs
Partition-Guided GANs
Check 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 |
|---|---|---|---|
| Add the partitioner guide term to Partition GAN (ablation) | Partition GAN · CIFAR-10 | FID: 10.69 ± 0.33 -> 8.93 ± 0.38 | Helped |
| Add the partitioner guide term to Partition GAN (ablation) | Partition GAN · CIFAR-10 | IS: 8.52 ± 0.075 -> 8.81 ± 0.101 | Helped |
| Add the partitioner guide term to Partition GAN (ablation) | Partition GAN · CIFAR-10 | reverse KL: 0.0005 ± 0.0002 -> 0.0004 ± 0.0001 | Helped |
| Replace random partition labels with learned partitions (ablation) | Partition GAN · CIFAR-10 | FID: 22.57 ± 0.59 -> 10.69 ± 0.33 | Helped |
| Replace random partition labels with learned partitions (ablation) | Partition GAN · CIFAR-10 | IS: 7.47 ± 0.051 -> 8.52 ± 0.075 | Helped |
| Replace random partition labels with learned partitions (ablation) | Partition GAN · CIFAR-10 | reverse KL: 0.0074 ± 0.0008 -> 0.0005 ± 0.0002 | Helped |
| Vary the number of partitions among 50, 100, and 200 (ablation) | PGMGAN · CIFAR-10 | FID: 9.27 ± 0.45; 8.97 ± 0.33; 8.93 ± 0.38 | No clear effect |
| Vary the number of partitions among 50, 100, and 200 (ablation) | PGMGAN · CIFAR-10 | IS: 8.69 ± 0.07; 8.75 ± 0.05; 8.81 ± 0.10 | No clear effect |
| Enforce the partitioner architecture constraints that prevent guide local optima (ablation) | PGMGAN · 2D-grid | recovered modes | Helped |
| Use the reliable guide instead of training GANs without a guide (ablation) | PGMGAN · 2D-grid | recovered modes | Helped |
| Use PGMGAN rather than Self-Cond-GAN | ImageNet 128×128 · ImageNet | FID: 40.30 -> 21.73 | Helped |
| Use PGMGAN rather than Self-Cond-GAN | ImageNet 128×128 · ImageNet | IS: 15.82 -> 23.31 | Helped |
Closest papers by meaning
- AdaGAN: Boosting Generative Models (NeurIPS 2017 · similarity 0.89)
- MGAN: Training Generative Adversarial Nets with Multiple Generators (ICLR 2018 · similarity 0.89)
- Mixture Density Generative Adversarial Networks (CVPR 2019 · similarity 0.89)
- A Large-Scale Study on Regularization and Normalization in GANs (ICML 2019 · similarity 0.89)
- Structure-preserving GANs (ICML 2022 · similarity 0.89)
- Mixed batches and symmetric discriminators for GAN training (ICML 2018 · similarity 0.89)
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.