ISDE : Independence Structure Density Estimation

TL;DR

This paper introduces ISDE, a density estimation algorithm that models multivariate data by exploiting independence structures to mitigate high-dimensional challenges, demonstrating competitive performance on synthetic and real data.

Contribution

ISDE is a novel density estimation method that explicitly incorporates independence structures, improving estimation efficiency in high-dimensional settings.

Findings

ISDE outperforms traditional methods in log-likelihood on synthetic data.

ISDE effectively identifies meaningful variable partitions.

The algorithm demonstrates reasonable complexity and runtime.

Abstract

In this paper, we propose ISDE (Independence Structure Density Estimation), an algorithm designed to estimate a multivariate density under Kullback-Leibler loss and the Independence Structure (IS) model. IS tackles the curse of dimensionality by separating features into independent groups. We explain the construction of ISDE and present some experiments to show its performance on synthetic and real-world data. Performance is measured quantitatively by comparing empirical $\log$-likelihood with other density estimation methods and qualitatively by analyzing outputted partitions of variables. We also provide information about complexity and running time.