On Probabilistic Embeddings in Optimal Dimension Reduction

TL;DR

This paper investigates the theoretical properties of probabilistic embeddings in dimension reduction, revealing that standard algorithms may produce non-deterministic solutions and that optimal solutions are deterministic, with implications for clustering and interpretability.

Contribution

It provides a theoretical analysis of a generalized multidimensional scaling problem, showing the difference between solutions from standard methods and the globally optimal deterministic embeddings.

Findings

Standard particle descent methods may produce non-deterministic embeddings.

Relaxed probabilistic formulations admit interpretable solutions.

Globally optimal solutions are deterministic and parametric.

Abstract

Dimension reduction algorithms are a crucial part of many data science pipelines, including data exploration, feature creation and selection, and denoising. Despite their wide utilization, many non-linear dimension reduction algorithms are poorly understood from a theoretical perspective. In this work we consider a generalized version of multidimensional scaling, which is posed as an optimization problem in which a mapping from a high-dimensional feature space to a lower-dimensional embedding space seeks to preserve either inner products or norms of the distribution in feature space, and which encompasses many commonly used dimension reduction algorithms. We analytically investigate the variational properties of this problem, leading to the following insights: 1) Solutions found using standard particle descent methods may lead to non-deterministic embeddings, 2) A relaxed or probabilistic formulation of the problem admits solutions with easily interpretable necessary conditions, 3) The globally optimal solutions to the relaxed problem actually must give a deterministic embedding. This progression of results mirrors the classical development of optimal transportation, and in a case relating to the Gromov-Wasserstein distance actually gives explicit insight into the structure of the optimal embeddings, which are parametrically determined and discontinuous. Finally, we illustrate that a standard computational implementation of this task does not learn deterministic embeddings, which means that it learns sub-optimal mappings, and that the embeddings learned in that context have highly misleading clustering structure, underscoring the delicate nature of solving this problem computationally.