Research map / Probabilistic and scientific modeling / Variational and generative sampling
Self-Adversarially Learned Bayesian Sampling
Yang Zhao, Jianyi Zhang, Changyou Chen
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 |
|---|---|---|---|
| use SAL-MC without an MH step instead of A-NICE-MC with MH | Bayesian sampling · Ring, Mog2, Mog6 | ESS/s: 121138 vs 43887; 72691 vs 17216; 65287 vs 11745 | Improved |
| use SAL-MC with MH instead of A-NICE-MC with MH | Bayesian sampling · Mog2, Mog6, Ring5 | ESS: 1172 vs 951; 978 vs 889; 414 vs 335 | Improved |
| add an MH step (ablation) | SAL-MC · Ring, Mog2, Mog6 | ESS/s: 121138 -> 25624; 72691 -> 17531; 65287 -> 12235 | Hurt |
| add an MH step (ablation) | SAL-MC · Heart, Australian, German | ESS: 1683 -> 1; 1385 -> 1; 1897 -> 1 | Hurt |
| use SAL-MC with MH instead of A-NICE-MC with MH | Bayesian logistic regression sampling · Heart, Australian, German | ESS: 1683 vs 663; 1385 vs 596; 1897 vs 483 | Improved |
| use SAL-MC instead of A-NICE-MC | Bayesian logistic regression · Heart, Australian, German | test accuracy: 84.10%, 88.38%, 80.32%; the same as A-NICE-MC | No clear effect |
| inject Gaussian noise with variance 0.5 instead of using zero noise (ablation) | learned generator sampling · MNIST | mixing across digit modes | Improved |
| inject noise that is too small or too large (ablation) | SAL-MC · synthetic multimodal distributions | effective sample size | Hurt |
| use AG-SVGD instead of SGLD | direct Bayesian sampling · 5D Gaussian mixture | variance estimation MSE | Improved |
| use AG-SVGD instead of A-SVGD | direct Bayesian sampling · 5D Gaussian mixture | sample approximation quality | No clear effect |
| use AG-SVGD instead of A-SVGD or SGLD | direct Bayesian sampling · 2D-GMM Mog4 | mode balance | Improved |
Closest papers by meaning
- Ai-sampler: Adversarial Learning of Markov kernels with involutive maps (ICML 2024 · similarity 0.87)
- A-NICE-MC: Adversarial Training for MCMC (NeurIPS 2017 · similarity 0.87)
- Microcanonical Langevin Ensembles: Advancing the Sampling of Bayesian Neural Networks (ICLR 2025 · similarity 0.85)
- Scalable Discrete Diffusion Samplers: Combinatorial Optimization and Statistical Physics (ICLR 2025 · similarity 0.84)
- MDNS: Masked Diffusion Neural Sampler via Stochastic Optimal Control (NeurIPS 2025 · similarity 0.84)
- Parameter Expanded Stochastic Gradient Markov Chain Monte Carlo (ICLR 2025 · similarity 0.84)
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.
