Learning disconnected manifolds: a no GANs land

06/08/2020
by   Ugo Tanielian, et al.
0

Typical architectures of Generative AdversarialNetworks make use of a unimodal latent distribution transformed by a continuous generator. Consequently, the modeled distribution always has connected support which is cumbersome when learning a disconnected set of manifolds. We formalize this problem by establishing a no free lunch theorem for the disconnected manifold learning stating an upper bound on the precision of the targeted distribution. This is done by building on the necessary existence of a low-quality region where the generator continuously samples data between two disconnected modes. Finally, we derive a rejection sampling method based on the norm of generators Jacobian and show its efficiency on several generators including BigGAN.

READ FULL TEXT

page 5

page 7

page 8

page 16

page 17

page 19

page 20

page 21

research
06/03/2018

Disconnected Manifold Learning for Generative Adversarial Networks

Real images often lie on a union of disjoint manifolds rather than one g...
research
07/21/2022

Optimal precision for GANs

When learning disconnected distributions, Generative adversarial network...
research
08/20/2021

MG-GAN: A Multi-Generator Model Preventing Out-of-Distribution Samples in Pedestrian Trajectory Prediction

Pedestrian trajectory prediction is challenging due to its uncertain and...
research
05/22/2021

Continuous Flattening of All Polyhedral Manifolds using Countably Infinite Creases

We prove that any finite polyhedral manifold in 3D can be continuously f...
research
07/01/2016

A scaled Bregman theorem with applications

Bregman divergences play a central role in the design and analysis of a ...
research
10/19/2021

Latent reweighting, an almost free improvement for GANs

Standard formulations of GANs, where a continuous function deforms a con...
research
02/02/2018

Selective Sampling and Mixture Models in Generative Adversarial Networks

In this paper, we propose a multi-generator extension to the adversarial...

Please sign up or login with your details

Forgot password? Click here to reset