Approximation of High-Dimensional Gibbs Distributions with Functional Hierarchical Tensors

Fuente: arXiv
Salvato in:
Dettagli Bibliografici
Autori principali: Sheng, Nan, Tang, Xun, Chen, Haoxuan, Ying, Lexing
Natura: Preprint
Pubblicazione: 2025
Soggetti:
Accesso online:
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866909469141630976
author Sheng, Nan
Tang, Xun
Chen, Haoxuan
Ying, Lexing
author_facet Sheng, Nan
Tang, Xun
Chen, Haoxuan
Ying, Lexing
contents The numerical representation of high-dimensional Gibbs distributions is challenging due to the curse of dimensionality manifesting through the intractable normalization constant calculations. This work addresses this challenge by performing a particle-based high-dimensional parametric density estimation subroutine, and the input to the subroutine is Gibbs samples generated by leveraging advanced sampling techniques. Specifically, to generate Gibbs samples, we employ ensemble-based annealed importance sampling, a population-based approach for sampling multimodal distributions. These samples are then processed using functional hierarchical tensor sketching, a tensor-network-based density estimation method for high-dimensional distributions, to obtain the numerical representation of the Gibbs distribution. We successfully apply the proposed approach to complex Ginzburg-Landau models with hundreds of variables. In particular, we show that the approach proposed is successful at addressing the metastability issue under difficult numerical cases.
format Preprint
id arxiv_https___arxiv_org_abs_2501_17143
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Approximation of High-Dimensional Gibbs Distributions with Functional Hierarchical Tensors
Sheng, Nan
Tang, Xun
Chen, Haoxuan
Ying, Lexing
Numerical Analysis
Computational Physics
Data Analysis, Statistics and Probability
The numerical representation of high-dimensional Gibbs distributions is challenging due to the curse of dimensionality manifesting through the intractable normalization constant calculations. This work addresses this challenge by performing a particle-based high-dimensional parametric density estimation subroutine, and the input to the subroutine is Gibbs samples generated by leveraging advanced sampling techniques. Specifically, to generate Gibbs samples, we employ ensemble-based annealed importance sampling, a population-based approach for sampling multimodal distributions. These samples are then processed using functional hierarchical tensor sketching, a tensor-network-based density estimation method for high-dimensional distributions, to obtain the numerical representation of the Gibbs distribution. We successfully apply the proposed approach to complex Ginzburg-Landau models with hundreds of variables. In particular, we show that the approach proposed is successful at addressing the metastability issue under difficult numerical cases.
title Approximation of High-Dimensional Gibbs Distributions with Functional Hierarchical Tensors
topic Numerical Analysis
Computational Physics
Data Analysis, Statistics and Probability
url https://arxiv.org/abs/2501.17143