AI Paper Atlas

Research map / Probabilistic and scientific modeling / Variational and generative sampling

Microcanonical Hamiltonian Monte Carlo

Jakob Robnik, G. Bruno De Luca, Eva Silverstein, Uroš Seljak

JMLR 2023 · Probabilistic and scientific modeling › Variational and generative sampling · approach: Stochastic gradient MCMC · 25 citations (OpenAlex)

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 changedWhereMeasuredEffect
use MCHMC (Langevin-like, q=0) instead of NUTSMCHMC · Ill-conditioned Gaussian kappa=100ESS: 0.012 -> 0.075Improved
use auto-tuned MCLMC instead of NUTS including tuning costMCLMC · Neal's funnelESS: 0.0019 -> 0.0078Improved
use auto-tuned MCLMC instead of NUTS including tuning costMCLMC · CauchyESS: < 10-6 -> 0.001Improved
use auto-tuned MCLMC instead of NUTSMCLMC · German creditESS: 0.0008 -> 0.0059Improved
use minimal norm integrator instead of leapfrog (ablation)MCHMC Langevin-like q=0 · Ill-conditioned GaussianESS: 0.075 -> 0.110Improved
use minimal norm integrator instead of leapfrog (ablation)MCHMC Langevin-like q=0 · German creditESS: 0.0105 -> 0.0099No clear effect
use q=0 Hamiltonian instead of q=2 (ablation)MCHMC · Neal's funnelESS: 0.0001 -> 0.021Improved
add momentum decoherence (bounces) to no-bounce MCHMC (ablation)MCHMC · Ill-conditioned GaussianESS: 0 -> 0.039Improved
use multiple chains (ESH) instead of single-chain no-bounce (ablation)MCHMC without decoherence · RosenbrockESS: 0.0012 -> 0.0004Hurt
use Langevin-like partial refresh instead of full bounces (ablation)MCHMC LF q=0 · Ill-conditioned GaussianESS: 0.039 -> 0.075Improved
use MCHMC (Langevin MN) instead of unadjusted HMCMCHMC · Stochastic VolatilityESS: 0.002 -> 0.023Improved
use unadjusted HMC instead of NUTSunadjusted HMC · RosenbrockESS: 0.0015 -> 0.0051Improved

Closest papers by meaning

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.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy