Research map / Probabilistic and scientific modeling / Variational and generative sampling
Stein $\Pi$-Importance Sampling
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 |
|---|---|---|---|
| Sample from Π ∝ p√kP instead of P before optimal Stein weighting | Stein importance sampling with the Langevin kernel · PosteriorDB tasks | KSD: significant improvement in 73% of test problems | Helped |
| Sample from Π ∝ p√kP instead of P before optimal Stein weighting | Stein importance sampling with the KGM3 kernel · PosteriorDB tasks | KSD: significant improvement in 65% of test problems | Helped |
| Use Stein post-processing rather than uniform weights on MALA samples | MALA posterior samples · PosteriorDB tasks with the Langevin kernel | KSD: significant improvement in 100% of cases | Helped |
| Use Stein post-processing rather than uniform weights on MALA samples | MALA posterior samples · PosteriorDB tasks with the KGM3 kernel | KSD: significant improvement in 97% of cases | Helped |
| Sample from Π rather than P before Stein weighting | Langevin Stein importance sampling · kidiq-kidscore_interaction | KSD: 1.65 -> 1.79 | Hurt |
| Sample from Π rather than P before Stein weighting | KGM3 Stein importance sampling · eight_schools-eight_schools_centered | KSD: 2.15 -> 215 | Hurt |
| Use the proposed Π rather than P for Stein importance sampling | Stein importance sampling · PosteriorDB tasks as dimension increases | KSD | Mixed |
| Use Π1 ∝ p^(d/(d+1)) rather than the proposed Π ∝ p√kP | Stein importance sampling · PosteriorDB, especially skewed targets | KSD | Hurt |
Closest papers by meaning
- Stein Point Markov Chain Monte Carlo (ICML 2019 · similarity 0.89)
- The Polynomial Stein Discrepancy for Assessing Moment Convergence (ICML 2025 · similarity 0.88)
- Estimating the Permanent by Nesting Importance Sampling (ICML 2024 · similarity 0.86)
- Multi-Marginal Couplings for Metropolis--Hastings (NeurIPS 2026 · similarity 0.85)
- Differentiable Annealed Importance Sampling Minimizes The Jensen-Shannon Divergence Between Initial and Target Distribution (ICML 2024 · similarity 0.85)
- Chain of Log-Concave Markov Chains (ICLR 2024 · 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.