A Spectral Framework for Closed-Form Relative Density Estimation

Fuente: arXiv
Saved in:
Bibliographic Details
Main Author: Bach, Francis
Format: Preprint
Published: 2026
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866914554225623040
author Bach, Francis
author_facet Bach, Francis
contents We propose a closed-form spectral framework for relative log-density estimation in linearly parameterized probabilistic models, including unnormalized and conditional models. This is achieved by representing the Kullback-Leibler (KL) divergence as an integral of weighted chi-squared divergences, converting KL estimation into a family of least-squares problems. We derive an explicit spectral formula based only on first- and second-order feature moments, yielding closed-form estimators of both divergences and log-density potentials for fixed features. The framework extends to a broad class of f-divergences and can be combined with kernelization or feature learning with neural networks. We prove convergence guarantees for the resulting estimators and empirically compare them on synthetic data with optimization-based variational formulations, including logistic and softmax regression for normalized conditional models.
format Preprint
id arxiv_https___arxiv_org_abs_2605_10668
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle A Spectral Framework for Closed-Form Relative Density Estimation
Bach, Francis
Machine Learning
Optimization and Control
Statistics Theory
We propose a closed-form spectral framework for relative log-density estimation in linearly parameterized probabilistic models, including unnormalized and conditional models. This is achieved by representing the Kullback-Leibler (KL) divergence as an integral of weighted chi-squared divergences, converting KL estimation into a family of least-squares problems. We derive an explicit spectral formula based only on first- and second-order feature moments, yielding closed-form estimators of both divergences and log-density potentials for fixed features. The framework extends to a broad class of f-divergences and can be combined with kernelization or feature learning with neural networks. We prove convergence guarantees for the resulting estimators and empirically compare them on synthetic data with optimization-based variational formulations, including logistic and softmax regression for normalized conditional models.
title A Spectral Framework for Closed-Form Relative Density Estimation
topic Machine Learning
Optimization and Control
Statistics Theory
url https://arxiv.org/abs/2605.10668