Comparing the latent space of generative models

TL;DR

This paper investigates the relationships between latent spaces of different generative models for human faces, revealing that simple linear mappings can effectively translate between them while preserving most information.

Contribution

It introduces a method to compare and transform latent spaces of different models, focusing on face data, and shows linear mappings are often sufficient for effective translation.

Findings

Linear mappings can effectively translate between different models' latent spaces.

Most information is preserved when transforming between models using simple linear mappings.

The approach is validated on generative models for human faces.

Abstract

Different encodings of datapoints in the latent space of latent-vector generative models may result in more or less effective and disentangled characterizations of the different explanatory factors of variation behind the data. Many works have been recently devoted to the explorationof the latent space of specific models, mostly focused on the study of how features are disentangled and of how trajectories producing desired alterations of data in the visible space can be found. In this work we address the more general problem of comparing the latent spaces of different models, looking for transformations between them. We confined the investigation to the familiar and largely investigated case of generative models for the data manifold of human faces. The surprising, preliminary result reported in this article is that (provided models have not been taught or explicitly conceived to act differently) a simple linear mapping is enough to pass from a latent space to another while preserving most of the information.