Research map / Probabilistic and scientific modeling / Variational and generative sampling
Microcanonical Hamiltonian Monte Carlo
Jakob Robnik, G. Bruno De Luca, Eva Silverstein, Uroš Seljak
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 MCHMC (Langevin-like, q=0) instead of NUTS | MCHMC · Ill-conditioned Gaussian kappa=100 | ESS: 0.012 -> 0.075 | Improved |
| use auto-tuned MCLMC instead of NUTS including tuning cost | MCLMC · Neal's funnel | ESS: 0.0019 -> 0.0078 | Improved |
| use auto-tuned MCLMC instead of NUTS including tuning cost | MCLMC · Cauchy | ESS: < 10-6 -> 0.001 | Improved |
| use auto-tuned MCLMC instead of NUTS | MCLMC · German credit | ESS: 0.0008 -> 0.0059 | Improved |
| use minimal norm integrator instead of leapfrog (ablation) | MCHMC Langevin-like q=0 · Ill-conditioned Gaussian | ESS: 0.075 -> 0.110 | Improved |
| use minimal norm integrator instead of leapfrog (ablation) | MCHMC Langevin-like q=0 · German credit | ESS: 0.0105 -> 0.0099 | No clear effect |
| use q=0 Hamiltonian instead of q=2 (ablation) | MCHMC · Neal's funnel | ESS: 0.0001 -> 0.021 | Improved |
| add momentum decoherence (bounces) to no-bounce MCHMC (ablation) | MCHMC · Ill-conditioned Gaussian | ESS: 0 -> 0.039 | Improved |
| use multiple chains (ESH) instead of single-chain no-bounce (ablation) | MCHMC without decoherence · Rosenbrock | ESS: 0.0012 -> 0.0004 | Hurt |
| use Langevin-like partial refresh instead of full bounces (ablation) | MCHMC LF q=0 · Ill-conditioned Gaussian | ESS: 0.039 -> 0.075 | Improved |
| use MCHMC (Langevin MN) instead of unadjusted HMC | MCHMC · Stochastic Volatility | ESS: 0.002 -> 0.023 | Improved |
| use unadjusted HMC instead of NUTS | unadjusted HMC · Rosenbrock | ESS: 0.0015 -> 0.0051 | Improved |
Closest papers by meaning
- Metropolis Adjusted Microcanonical Hamiltonian Monte Carlo (NeurIPS 2025 · similarity 0.90)
- Magnetic Hamiltonian Monte Carlo (ICML 2017 · similarity 0.90)
- Reflective Hamiltonian Monte Carlo: Mixing Analysis and Application to Sampling on Stiefel Manifold (ICML 2026 · similarity 0.88)
- Practical and Scalable Hamiltonian Monte Carlo Without the Metropolis Test (ICML 2026 · similarity 0.87)
- Hamiltonian Monte Carlo with Energy Conserving Subsampling (JMLR 2019 · similarity 0.87)
- Practical Hamiltonian Monte Carlo on Riemannian Manifolds via Relativity Theory (ICML 2024 · 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.
