On variational inference and maximum likelihood estimation with the λ-exponential family

Fuente: arXiv
Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Guilmeau, Thomas, Chouzenoux, Emilie, Elvira, Víctor
Format: Preprint
Veröffentlicht: 2023
Schlagworte:
Online-Zugang:
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866910494828265472
author Guilmeau, Thomas
Chouzenoux, Emilie
Elvira, Víctor
author_facet Guilmeau, Thomas
Chouzenoux, Emilie
Elvira, Víctor
contents The λ-exponential family has recently been proposed to generalize the exponential family. While the exponential family is well-understood and widely used, this it not the case of the λ-exponential family. However, many applications require models that are more general than the exponential family. In this work, we propose a theoretical and algorithmic framework to solve variational inference and maximum likelihood estimation problems over the λ-exponential family. We give new sufficient optimality conditions for variational inference problems. Our conditions take the form of generalized moment-matching conditions and generalize existing similar results for the exponential family. We exhibit novel characterizations of the solutions of maximum likelihood estimation problems, that recover optimality conditions in the case of the exponential family. For the resolution of both problems, we propose novel proximal-like algorithms that exploit the geometry underlying the λ-exponential family. These new theoretical and methodological insights are tested on numerical examples, showcasing their usefulness and interest, especially on heavy-tailed target distributions.
format Preprint
id arxiv_https___arxiv_org_abs_2310_05781
institution arXiv
publishDate 2023
record_format arxiv
spellingShingle On variational inference and maximum likelihood estimation with the λ-exponential family
Guilmeau, Thomas
Chouzenoux, Emilie
Elvira, Víctor
Statistics Theory
62F99, 62B11, 49K10, 90C26
The λ-exponential family has recently been proposed to generalize the exponential family. While the exponential family is well-understood and widely used, this it not the case of the λ-exponential family. However, many applications require models that are more general than the exponential family. In this work, we propose a theoretical and algorithmic framework to solve variational inference and maximum likelihood estimation problems over the λ-exponential family. We give new sufficient optimality conditions for variational inference problems. Our conditions take the form of generalized moment-matching conditions and generalize existing similar results for the exponential family. We exhibit novel characterizations of the solutions of maximum likelihood estimation problems, that recover optimality conditions in the case of the exponential family. For the resolution of both problems, we propose novel proximal-like algorithms that exploit the geometry underlying the λ-exponential family. These new theoretical and methodological insights are tested on numerical examples, showcasing their usefulness and interest, especially on heavy-tailed target distributions.
title On variational inference and maximum likelihood estimation with the λ-exponential family
topic Statistics Theory
62F99, 62B11, 49K10, 90C26
url https://arxiv.org/abs/2310.05781