Research map / Probabilistic and scientific modeling / Variational and generative sampling
Bridging Explicit and Implicit Deep Generative Models via Neural Stein Estimators
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 |
|---|---|---|---|
| joint training with explicit energy model via Stein bridging vs WGAN-GP | Joint-W · CIFAR-10 | Inception Score: 5.6% | Helped |
| joint training vs WGAN-GP | Joint-W · CIFAR-10 | FID | Helped |
| joint training vs entropy-regularized WGAN (variational annealing) | Joint-W · CIFAR-10 | FID | Helped |
| joint training vs GAN, WGAN-GP, DGM | Joint-JS and Joint-W · Two-Circle and Two-Spiral | MMD and HSR | Helped |
| joint training reduces metric variance during training | Joint-W · MNIST | learning-curve variance | Helped |
| joint training reduces model distribution bias | Joint-W · MNIST and CIFAR-10 | mean distance to true data | Helped |
| noisy training data | Joint-W vs WGAN-GP · Two-Circle | sample quality | Helped |
| reduce training sample size from 2000 to 100 | DEM · Two-Spiral | AUC: 2000 -> 100 | Hurt |
| reduce training sample size below 500 | Joint-W · Two-Spiral | AUC: 500 | Mixed |
| DEM and EGAN on OOD detection | DEM and EGAN · CIFAR-10 OOD | AUC | Hurt |
| Joint-W on OOD detection | Joint-W · CIFAR-10 OOD | AUC | Helped |
Closest papers by meaning
- Bias and Generalization in Deep Generative Models: An Empirical Study (NeurIPS 2018 · similarity 0.89)
- Auxiliary Deep Generative Models (ICML 2016 · similarity 0.88)
- Stein Transport for Generative Modeling (NeurIPS 2026 · similarity 0.88)
- A Likelihood Based Approach to Distribution Regression Using Conditional Deep Generative Models (ICML 2025 · similarity 0.87)
- Statistically Optimal Generative Modeling with Maximum Deviation from the Empirical Distribution (ICML 2024 · similarity 0.86)
- On the Stability of Iterative Retraining of Generative Models on their own Data (ICLR 2024 · similarity 0.86)
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.