Parameter estimation for Gibbs distributions

Fuente: arXiv
Guardado en:
Detalles Bibliográficos
Autores principales: Harris, David G., Kolmogorov, Vladimir
Formato: Preprint
Publicado: 2020
Materias:
Acceso en línea:
Etiquetas: Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
_version_ 1866909562278248448
author Harris, David G.
Kolmogorov, Vladimir
author_facet Harris, David G.
Kolmogorov, Vladimir
contents We consider Gibbs distributions, which are families of probability distributions over a discrete space $Ω$ with probability mass function of the form $μ^Ω_β(ω) \propto e^{βH(ω)}$ for $β$ in an interval $[β_{\min}, β_{\max}]$ and $H( ω) \in \{0 \} \cup [1, n]$. The partition function is the normalization factor $Z(β)=\sum_{ω\inΩ}e^{βH(ω)}$. Two important parameters of these distributions are the log partition ratio $q = \log \tfrac{Z(β_{\max})}{Z(β_{\min})}$ and the counts $c_x = |H^{-1}(x)|$. These are correlated with system parameters in a number of physical applications and sampling algorithms. Our first main result is to estimate the counts $c_x$ using roughly $\tilde O( \frac{q}{\varepsilon^2})$ samples for general Gibbs distributions and $\tilde O( \frac{n^2}{\varepsilon^2} )$ samples for integer-valued distributions (ignoring some second-order terms and parameters), and we show this is optimal up to logarithmic factors. We illustrate with improved algorithms for counting connected subgraphs, independent sets, and perfect matchings. As a key subroutine, we also develop algorithms to compute the partition function $Z$ using $\tilde O(\frac{q}{\varepsilon^2})$ samples for general Gibbs distributions and using $\tilde O(\frac{n^2}{\varepsilon^2})$ samples for integer-valued distributions.
format Preprint
id arxiv_https___arxiv_org_abs_2007_10824
institution arXiv
publishDate 2020
record_format arxiv
spellingShingle Parameter estimation for Gibbs distributions
Harris, David G.
Kolmogorov, Vladimir
Data Structures and Algorithms
Computational Complexity
Discrete Mathematics
Probability
We consider Gibbs distributions, which are families of probability distributions over a discrete space $Ω$ with probability mass function of the form $μ^Ω_β(ω) \propto e^{βH(ω)}$ for $β$ in an interval $[β_{\min}, β_{\max}]$ and $H( ω) \in \{0 \} \cup [1, n]$. The partition function is the normalization factor $Z(β)=\sum_{ω\inΩ}e^{βH(ω)}$. Two important parameters of these distributions are the log partition ratio $q = \log \tfrac{Z(β_{\max})}{Z(β_{\min})}$ and the counts $c_x = |H^{-1}(x)|$. These are correlated with system parameters in a number of physical applications and sampling algorithms. Our first main result is to estimate the counts $c_x$ using roughly $\tilde O( \frac{q}{\varepsilon^2})$ samples for general Gibbs distributions and $\tilde O( \frac{n^2}{\varepsilon^2} )$ samples for integer-valued distributions (ignoring some second-order terms and parameters), and we show this is optimal up to logarithmic factors. We illustrate with improved algorithms for counting connected subgraphs, independent sets, and perfect matchings. As a key subroutine, we also develop algorithms to compute the partition function $Z$ using $\tilde O(\frac{q}{\varepsilon^2})$ samples for general Gibbs distributions and using $\tilde O(\frac{n^2}{\varepsilon^2})$ samples for integer-valued distributions.
title Parameter estimation for Gibbs distributions
topic Data Structures and Algorithms
Computational Complexity
Discrete Mathematics
Probability
url https://arxiv.org/abs/2007.10824