Convergence rates for random feature neural network approximation in molecular dynamics

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Huang, Xin, Plechac, Petr, Sandberg, Mattias, Szepessy, Anders
Format: Preprint
Published: 2024
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866910497006157824
author Huang, Xin
Plechac, Petr
Sandberg, Mattias
Szepessy, Anders
author_facet Huang, Xin
Plechac, Petr
Sandberg, Mattias
Szepessy, Anders
contents Random feature neural network approximations of the potential in Hamiltonian systems yield approximations of molecular dynamics correlation observables that have the expected error $\mathcal{O}\big((K^{-1}+J^{-1/2})^{\frac{1}{2}}\big)$, for networks with $K$ nodes using $J$ data points, provided the Hessians of the potential and the observables are bounded. The loss function is based on the least squares error of the potential and regularizations, with the data points sampled from the Gibbs density. The proof uses an elementary new derivation of the generalization error for random feature networks that does not apply the Rademacher or related complexities.
format Preprint
id arxiv_https___arxiv_org_abs_2406_14791
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Convergence rates for random feature neural network approximation in molecular dynamics
Huang, Xin
Plechac, Petr
Sandberg, Mattias
Szepessy, Anders
Numerical Analysis
82C32, 82M31, 65K10, 65P10
Random feature neural network approximations of the potential in Hamiltonian systems yield approximations of molecular dynamics correlation observables that have the expected error $\mathcal{O}\big((K^{-1}+J^{-1/2})^{\frac{1}{2}}\big)$, for networks with $K$ nodes using $J$ data points, provided the Hessians of the potential and the observables are bounded. The loss function is based on the least squares error of the potential and regularizations, with the data points sampled from the Gibbs density. The proof uses an elementary new derivation of the generalization error for random feature networks that does not apply the Rademacher or related complexities.
title Convergence rates for random feature neural network approximation in molecular dynamics
topic Numerical Analysis
82C32, 82M31, 65K10, 65P10
url https://arxiv.org/abs/2406.14791