Research map / Probabilistic and scientific modeling / Variational and generative sampling
On Markov chain Monte Carlo methods for tall data
Rémi Bardenet, Arnaud Doucet, Chris Holmes
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 confidence MH sampler with a single Taylor proxy instead of full-data MH | Metropolis-Hastings · Synthetic 2-D logistic regression | likelihood evaluations per iteration: fraction of evaluations decreases by a factor 10 when n is multiplied by 10; 1 000 points suffice | Improved |
| use confidence sampler with Taylor proxies dropped every 10 iterations instead of full-data MH | Metropolis-Hastings · covtype.binary logistic regression | likelihood evaluations per iteration (% of n): 27 to 42% of n | Improved |
| recompute the proxy less often than every 10 iterations (ablation) | Confidence sampler with Taylor proxies · covtype.binary logistic regression | average likelihood evaluations per iteration: 2x10=20% of n due to recomputing the proxy at alpha=10 | Mixed |
Closest papers by meaning
- Scalable Metropolis-Hastings for Exact Bayesian Inference with Large Datasets (ICML 2019 · similarity 0.91)
- Spectral Subsampling MCMC for Stationary Time Series (ICML 2020 · similarity 0.88)
- Hamiltonian Monte Carlo with Energy Conserving Subsampling (JMLR 2019 · similarity 0.88)
- Markov Chain Monte Carlo without Evaluating the Target: an Auxiliary Variable Approach (ICML 2026 · similarity 0.88)
- Multi-Marginal Couplings for Metropolis--Hastings (NeurIPS 2026 · similarity 0.87)
- Parallelizing MCMC Across the Sequence Length (NeurIPS 2025 · similarity 0.86)
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.
