analiza aiqbench
This research introduces Neural Non-Equilibrium Hamiltonian Monte Carlo (NHMC), a novel framework that combines neural network training with statistical mechanics principles to sample from unnormalized Boltzmann densities. By leveraging non-equilibrium work and path-space importance sampling, the method aims to improve efficiency in complex physical systems like lattice gauge theories and molecular dynamics, though it faces challenges when proposal paths have insufficient overlap with the target distribution.
Neural Non-Equilibrium Hamiltonian Monte Carlo for Corrected Boltzmann Sampling
arXiv:2607.15682v2 Announce Type: replace Abstract: Sampling from an unnormalized Boltzmann density requires proposals that move probability mass globally while retaining enough path-probability information for statistical correction. We introduce Neural Non-Equilibrium Hamiltonian Monte Carlo (NHMC), a train-then-correct learned Hamiltonian sampler. Starting from a tractable base distribution, NHMC learns stochastic Hamiltonian-style paths toward the target. Once training is complete, the learned proposal parameters are fixed; the proposal then generates complete paths and endpoint configurations, which are statistically corrected using the recorded non-equilibrium work. This dimensionless generalized work is determined by the probability ratio between the forward proposal path and a reverse reference path. During training, minimizing its mean reduces a path-space KL divergence and controls an upper bound on endpoint mismatch. During evaluation, the same quantity defines weights for self-normalized importance sampling on paths (path-SNIS), estimates normalizing constants or free-energy differences, and gives the acceptance ratio for path-space independent Metropolis--Hastings (path-IMH). We further derive a shared-bridge round-trip NHMC--MH kernel and prove that its configuration-space transition preserves the Boltzmann target. On double-well, finite-volume lattice $\phi^4$, compact non-Abelian gauge, and Lennard--Jones cluster targets, the NHMC construction gives corrected estimates when path overlap is sufficient; when overlap is poor, weight degeneracy, low acceptance, and long autocorrelation expose proposal failure. We additionally report a molecular internal-coordinate feasibility study using a molecular-dynamics prior and learned-force path proposal.
Abstract:Sampling from an unnormalized Boltzmann density requires proposals that move probability mass globally while retaining enough path-probability information for statistical correction. We introduce Neural Non-Equilibrium Hamiltonian Monte Carlo (NHMC), a train-then-correct learned Hamiltonian sampler. Starting from a tractable base distribution, NHMC learns stochastic Hamiltonian-style paths toward the target. Once training is complete, the learned proposal parameters are fixed; the proposal then generates complete paths and endpoint configurations, which are statistically corrected using the recorded non-equilibrium work. This dimensionless generalized work is determined by the probability ratio between the forward proposal path and a reverse reference path. During training, minimizing its mean reduces a path-space KL divergence and controls an upper bound on endpoint mismatch. During evaluation, the same quantity defines weights for self-normalized importance sampling on paths (path-SNIS), estimates normalizing constants or free-energy differences, and gives the acceptance ratio for path-space independent Metropolis--Hastings (path-IMH). We further derive a shared-bridge round-trip NHMC--MH kernel and prove that its configuration-space transition preserves the Boltzmann target. On double-well, finite-volume lattice $\phi^4$, compact non-Abelian gauge, and Lennard--Jones cluster targets, the NHMC construction gives corrected estimates when path overlap is sufficient; when overlap is poor, weight degeneracy, low acceptance, and long autocorrelation expose proposal failure. We additionally report a molecular internal-coordinate feasibility study using a molecular-dynamics prior and learned-force path proposal.
Machine Learning (cs.LG); Statistical Mechanics (cond-mat.stat-mech); High Energy Physics - Lattice (hep-lat)
arXiv:2607.15682 [cs.LG]
(or arXiv:2607.15682v2 [cs.LG] for this version)
https://doi.org/10.48550/arXiv.2607.15682 Focus to learn more arXiv-issued DOI via DataCite
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Text integral sub licența deschisă a publicației · 316 cuvinte
Subiecte conexe
Neural Non-Equilibrium Hamiltonian Monte Carlo for Corrected Boltzmann Sampling
arXiv · 20 iul. 2026
Gibbs state preparation for commuting Hamiltonian: Mapping to classical Gibbs sampling
Quantum Journal · 18 sept. 2026
Diffusion-based Annealed Boltzmann Generators : benefits, pitfalls and hopes
arXiv · 23 iul. 2026
Learning the Universe: Posterior Reliability of Neural Generative Models in High-Dimensional Field-Level Inference of Cosmic Initial Conditions
arXiv · 10 iun. 2026