Research map / Probabilistic and scientific modeling / Variational and generative sampling
Markov Chain Monte Carlo without Evaluating the Target: an Auxiliary Variable Approach
Wei Yuan, Guanyang Wang
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 |
|---|---|---|---|
| add gradient-informed Barker proposal using minibatch gradient estimate | PoissonMH · Heterogeneous truncated Gaussian | best ESS/s (min): 0.40 -> 0.91 | Improved |
| add gradient-informed MALA proposal using minibatch gradient estimate | PoissonMH · Heterogeneous truncated Gaussian | best ESS/s (max): 4.67 -> 23.84 | Improved |
| add gradient-informed Barker proposal using minibatch gradient estimate | PoissonMH · Robust linear regression | best ESS/s (median): 100.2 -> 268.3 | Improved |
| add gradient-informed MALA proposal using minibatch gradient estimate | PoissonMH · Robust linear regression | best ESS/s (median): 100.2 -> 491.8 | Improved |
| use minibatch exact Poisson--Barker instead of full-batch Barker | Barker · Heterogeneous truncated Gaussian | best ESS/s (median): 0.22 -> 1.65 | Improved |
| use Poisson--Barker instead of random-walk Metropolis | MH · Heterogeneous truncated Gaussian | best ESS/s (median): 0.08 -> 1.65 | Improved |
| use Poisson--MALA instead of PoissonMH at low acceptance rate 0.25 | PoissonMH · Heterogeneous truncated Gaussian | MSE over time | Hurt |
| add minibatch Metropolis correction to SGLD proposal (Tuna--SGLD vs SGLD) | SGLD · Truncated Gaussian correctness check | max KS statistic: 0.18 -> 0.05 | Improved |
| add SGLD gradient-based proposal on a second minibatch | TunaMH · MNIST 3 vs 5 logistic regression | test accuracy over time | Improved |
| use full-batch methods instead of minibatch Poisson methods | Poisson--MALA · Robust linear regression | best ESS/s: 491.8 -> 5.1 | Hurt |
Closest papers by meaning
- Pseudo-Extended Markov chain Monte Carlo (NeurIPS 2019 · similarity 0.89)
- On Markov chain Monte Carlo methods for tall data (JMLR 2017 · similarity 0.88)
- Auxiliary Variational MCMC (ICLR 2019 · similarity 0.87)
- Involutive MCMC: a Unifying Framework (ICML 2020 · similarity 0.87)
- Multi-Marginal Couplings for Metropolis--Hastings (NeurIPS 2026 · similarity 0.86)
- Parallelizing MCMC Across the Sequence Length (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.
