DeepAI AI Chat
Log In Sign Up

Identifying latent distances with Finslerian geometry

12/20/2022
by   Alison Pouplin, et al.
University of Cambridge
RISE Research Institutes of Sweden
DTU
0

Riemannian geometry provides powerful tools to explore the latent space of generative models while preserving the inherent structure of the data manifold. Lengths, energies and volume measures can be derived from a pullback metric, defined through the immersion that maps the latent space to the data space. With this in mind, most generative models are stochastic, and so is the pullback metric. Manipulating stochastic objects is strenuous in practice. In order to perform operations such as interpolations, or measuring the distance between data points, we need a deterministic approximation of the pullback metric. In this work, we are defining a new metric as the expected length derived from the stochastic pullback metric. We show this metric is Finslerian, and we compare it with the expected pullback metric. In high dimensions, we show that the metrics converge to each other at a rate of 𝒪(1/D).

READ FULL TEXT

page 8

page 9

page 10

page 11

page 15

page 16

03/09/2021

A prior-based approximate latent Riemannian metric

Stochastic generative models enable us to capture the geometric structur...
08/20/2019

Expected path length on random manifolds

Manifold learning seeks a low dimensional representation that faithfully...
08/02/2020

Geometrically Enriched Latent Spaces

A common assumption in generative models is that the generator immerses ...
02/12/2020

Learning Flat Latent Manifolds with VAEs

Measuring the similarity between data points often requires domain knowl...
03/27/2020

LIMP: Learning Latent Shape Representations with Metric Preservation Priors

In this paper, we advocate the adoption of metric preservation as a powe...
06/09/2021

Pulling back information geometry

Latent space geometry has shown itself to provide a rich and rigorous fr...
10/22/2020

Geometry-Aware Hamiltonian Variational Auto-Encoder

Variational auto-encoders (VAEs) have proven to be a well suited tool fo...