Research map / Probabilistic and scientific modeling / Variational and generative sampling
Precision-Recall Divergence Optimization for Generative Modeling with GANs and Normalizing Flows
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 |
|---|---|---|---|
| Fine-tune BigGAN with PR-divergence λ=0.1 | BigGAN · CIFAR-10 32 × 32 | Recall: 65.66 -> 74.58 | Helped |
| Fine-tune BigGAN with PR-divergence λ=0.1 | BigGAN · CIFAR-10 32 × 32 | FID: 13.37 -> 11.62 | Helped |
| Fine-tune BigGAN with PR-divergence λ=20.0 | BigGAN · CIFAR-10 32 × 32 | Precision: 86.51 -> 90.03 | Helped |
| Fine-tune BigGAN with PR-divergence λ=0.5 | BigGAN · CelebA 64 × 64 | Recall: 51.42 -> 55.98 | Helped |
| Fine-tune BigGAN with PR-divergence λ=0.5 | BigGAN · CelebA 64 × 64 | FID: 9.16 -> 6.03 | Helped |
| Fine-tune BigGAN with PR-divergence λ=5.0 | BigGAN · CelebA 64 × 64 | Precision: 78.41 -> 83.96 | Helped |
| Fine-tune BigGAN with PR-divergence λ=5.0 | BigGAN · ImageNet 128 × 128 | Precision: 28.04 -> 30.87 | Helped |
| Fine-tune BigGAN with PR-divergence λ=0.5 | BigGAN · ImageNet 128 × 128 | Recall: 41.21 -> 42.38 | Helped |
| Fine-tune BigGAN with PR-divergence λ=5.0 | BigGAN · FFHQ 256 × 256 | Precision: 65.57 -> 84.31 | Helped |
| Fine-tune BigGAN with PR-divergence λ=0.5 | BigGAN · FFHQ 256 × 256 | Recall: 10.17 -> 9.66 | Hurt |
| Use hard truncation instead of PR-divergence fine-tuning | Baseline BigGAN · CIFAR-10 32 × 32 | Precision: 86.51 -> 88.03; 65.66 -> 53.63 | Mixed |
| Train the discriminator with auxiliary χ² instead of KL (ablation) | PR-divergence training · Synthetic 2D mixtures | Primal divergence estimation error | Helped |
Closest papers by meaning
- Assessing Generative Models via Precision and Recall (NeurIPS 2018 · similarity 0.92)
- Large Scale GAN Training for High Fidelity Natural Image Synthesis (ICLR 2019 · similarity 0.89)
- A Large-Scale Study on Regularization and Normalization in GANs (ICML 2019 · similarity 0.89)
- Statistically Optimal Generative Modeling with Maximum Deviation from the Empirical Distribution (ICML 2024 · similarity 0.87)
- Consistency-GAN: Training GANs with Consistency Model (AAAI 2024 · similarity 0.87)
- MCGAN: Enhancing GAN Training with Regression-Based Generator Loss (AAAI 2025 · 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.