Finite-Rank Logistic Gaussian Processes with Exact Likelihood for Conditional Density Estimation

2026-10-07

AI Summary

This study addresses the difficulty of conditional density estimation in logistic Gaussian processes (LGPs) arising from the absence of a closed-form normalizing constant. We propose a finite-rank LGP model that decomposes the log-density into a sum of bivariate functions endowed with piecewise-linear Gaussian priors. This formulation achieves, for the first time, a finite-rank LGP construction with an exact likelihood, yielding closed-form expressions for the normalizing constant, mean, and quantiles, while establishing minimax convergence rates. Efficient posterior inference is performed by combining Gibbs sampling with elliptical slice sampling. Experiments on synthetic data validate the proposed method, and its application to Alzheimer’s disease neuroimaging analysis successfully reveals the effect of diagnosis on the distribution of fractional anisotropy in the corpus callosum.

Abstract

Conditional density estimation describes how the entire distribution of a response changes with covariates, and in imaging studies also with location. Logistic Gaussian processes (LGP) give a flexible prior for such densities. However, the normalizing constant of an LGP has no closed form, so existing methods approximate or replace the likelihood. We propose the exact likelihood finite-rank LGP (ExFR-LGP), which writes the log density as the sum of two bivariate functions, one of the response and a location-varying linear index of the covariates, and one of the response and the location. Each function is assigned a finite-rank Gaussian process prior that is piecewise linear on a regular grid, enabling the normalizing constant, the conditional mean and the quantiles to enjoy closed form representations. Posterior samples are drawn from the exact posterior under this prior via a Gibbs sampler that updates the Gaussian components by elliptical slice sampling. When the true log density is the sum of two such bivariate functions, we show that the posterior contracts at the minimax rate of an $α$-smooth bivariate density up to a logarithmic factor. Numerical experiments using synthetic data demonstrate the effectiveness of ExFR-LGP in conditional density estimation. Furthermore, application to fractional anisotropy responses along the corpus callosum in the Alzheimer's Disease Neuroimaging Initiative data yields covariate-adjusted percentile bands with uncertainty, elucidating how diagnosis changes the distribution along the tract.

Problem

• Conditional Density Estimation
• Logistic Gaussian Processes
• Exact Likelihood
• Normalizing Constant

Innovation

• Conditional Density Estimation
• Logistic Gaussian Processes
• Finite-Rank Approximation
• Exact Likelihood
• Elliptical Slice Sampling