Latent Space Non-Linear Statistics

TL;DR

This paper introduces methods for performing statistical analysis directly on the nonlinear Riemannian latent space of deep generative models, enabling more accurate data interpretation beyond linear assumptions.

Contribution

It develops new techniques for nonlinear manifold statistics in latent spaces, including maximum likelihood inference and neural network approximations of geometric tensors.

Findings

Effective maximum likelihood inference in latent space.

Neural network approximation of Riemannian metric tensors.

Enhanced statistical analysis of deep generative models.

Abstract

Given data, deep generative models, such as variational autoencoders (VAE) and generative adversarial networks (GAN), train a lower dimensional latent representation of the data space. The linear Euclidean geometry of data space pulls back to a nonlinear Riemannian geometry on the latent space. The latent space thus provides a low-dimensional nonlinear representation of data and classical linear statistical techniques are no longer applicable. In this paper we show how statistics of data in their latent space representation can be performed using techniques from the field of nonlinear manifold statistics. Nonlinear manifold statistics provide generalizations of Euclidean statistical notions including means, principal component analysis, and maximum likelihood fits of parametric probability distributions. We develop new techniques for maximum likelihood inference in latent space, and adress the computational complexity of using geometric algorithms with high-dimensional data by training a separate neural network to approximate the Riemannian metric and cometric tensor capturing the shape of the learned data manifold.