Research map / Probabilistic and scientific modeling / Uncertainty and Gaussian processes
Scalable Nonparametric Sampling from Multimodal Posteriors with the Posterior Bootstrap
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 |
|---|---|---|---|
| Use posterior-bootstrap sampling with random restarts | NPL on Toy GMM · Toy GMM | posterior mode coverage | Helped |
| Use RR-NPL instead of NUTS | Toy GMM · Toy GMM | LPPD: -1.908 -> -1.909 | No clear effect |
| Use RR-NPL instead of NUTS | Toy GMM · Toy GMM | run-time for 2000 samples: 1m20s ± 16s -> 37.2s ± 4.5s | Helped |
| Use FI-NPL instead of ADVI | MNIST GMM · MNIST | LPPD: 1188.2 -> 2463.4 ±24.1 | Helped |
| Use Loss-NPL instead of ADVI | sparse logistic regression · Adult | LPPD: -0.327 ± 0.004 -> -0.326±0.004 | No clear effect |
| Use Loss-NPL instead of ADVI | sparse logistic regression · Polish | LPPD: -0.247 ± 0.047 -> -0.229 ± 0.034 | Helped |
| Use Loss-NPL instead of ADVI | sparse logistic regression · Arcene | LPPD: -0.445 ±0.068 -> -0.449 ± 0.104 | No clear effect |
| Use Loss-NPL instead of NUTS | sparse logistic regression · Arcene | LPPD: -0.464 ± 0.032 -> -0.449 ± 0.104 | Helped |
| Use Loss-NPL instead of NUTS | sparse logistic regression · Polish | LPPD: -3.336± 4.162 -> -0.229 ± 0.034 | Helped |
| Increase random restarts from R=1 to R=10 (ablation) | RR-NPL on toy separable GMM · Toy separable GMM | posterior mode representation | Helped |
| Increase random restarts from R=1 to R=10 (ablation) | RR-NPL on toy separable GMM · Toy separable GMM | run-time for 2000 posterior samples: 4.9 ± 0.6 -> 37.2s ± 4.5s | Hurt |
| Use fixed initialization rather than random initialization (ablation) | NPL on GMM · Toy GMM | label-switching | Helped |
Closest papers by meaning
- Scalable Bayesian Learning with posteriors (ICLR 2025 · similarity 0.88)
- Enhancing Transfer Learning with Flexible Nonparametric Posterior Sampling (ICLR 2024 · similarity 0.87)
- Provable Posterior Sampling with Denoising Oracles via Tilted Transport (NeurIPS 2024 · similarity 0.86)
- Improving Gradient-Guided Nested Sampling for Posterior Inference (ICML 2024 · similarity 0.86)
- Divide-and-Conquer Posterior Sampling for Denoising Diffusion priors (NeurIPS 2024 · similarity 0.86)
- Variational Inference with Mixtures of Isotropic Gaussians (NeurIPS 2025 · similarity 0.85)
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.