Research map / Probabilistic and scientific modeling / Variational and generative sampling
Measuring the reliability of MCMC inference with bidirectional Monte Carlo
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 BREAD's bidirectional AIS bound to estimate Jeffreys divergence | AIS on EASY RANDOM · EASY RANDOM, 1000 intermediate distributions | Jeffreys divergence: 0.00518 vs 0.0705 | Mixed |
| Use BREAD's bidirectional AIS bound to estimate Jeffreys divergence | AIS on HARD RANDOM · HARD RANDOM, 1000 intermediate distributions | Jeffreys divergence: 1.840 vs 2.309 | Mixed |
| Use BREAD's bidirectional AIS bound to estimate Jeffreys divergence | AIS on BARRIER · BARRIER, 1000 intermediate distributions | Jeffreys divergence: 1.085 vs 1.184 | Mixed |
| Collapse the latent matrix U | Matrix factorization in Stan · Matrix factorization (N=50, K=5, D=25) | convergence per MCMC step | Helped |
| Collapse the latent matrix U | Matrix factorization in Stan · Matrix factorization (N=50, K=5, D=25) | convergence per running time | Hurt |
| Collapse the latent matrix U | Matrix factorization in WebPPL · Matrix factorization (N=10, K=5, D=10) | convergence per running time | Helped |
| Initialize reverse AIS from fixed true hyperparameters and run a small number of MCMC steps (ablation) | BREAD fixed-hyperparameter scheme · Bayesian regression with broad Cauchy hyperparameter priors | stochastic upper bound | No clear effect |
| Use simulated data matched to hyperparameters inferred from real data | BREAD on Bayesian linear regression · NHEFS and simulated dataset | rate of change in forward AIS lower bounds | No clear effect |
Closest papers by meaning
- Surrogate Likelihoods for Variational Annealed Importance Sampling (ICML 2022 · similarity 0.90)
- A Divergence Bound for Hybrids of MCMC and Variational Inference and an Application to Langevin Dynamics and SGVI (ICML 2017 · similarity 0.89)
- Pseudo-Extended Markov chain Monte Carlo (NeurIPS 2019 · similarity 0.89)
- Involutive MCMC: a Unifying Framework (ICML 2020 · similarity 0.89)
- Parallelizing MCMC Across the Sequence Length (NeurIPS 2025 · similarity 0.88)
- Microcanonical Langevin Ensembles: Advancing the Sampling of Bayesian Neural Networks (ICLR 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.