Mssgan: Enforcing Multiple Generators to Learn Multiple Subspaces to Avoid the Mode Collapse

Miguel S. Soriano-García

doi:10.3390/make5040073

Mssgan: Enforcing Multiple Generators to Learn Multiple Subspaces to Avoid the Mode Collapse

2023-10-10

0

OA

AI

M

Miguel S. Soriano-García *

DOI：10.3390/make5040073

Original

Original request for help

Save

Abstract

En 中文

Generative Adversarial Networks are powerful generative models that are used in different areas and with multiple applications. However, this type of model has a training problem called mode collapse. This problem causes the generator to not learn the complete distribution of the data with which it is trained. To force the network to learn the entire data distribution, MSSGAN is introduced. This model has multiple generators and distributes the training data in multiple subspaces, where each generator is enforced to learn only one of the groups with the help of a classifier. We demonstrate that our model performs better on the FID and Sample Distribution metrics compared to previous models to avoid mode collapse. Experimental results show how each of the generators learns different information and, in turn, generates satisfactory quality samples.

Keywords:

generative models

mode collapse

GAN

multiple generators

Mssgan: Enforcing Multiple Generators to Learn Multiple Subspaces to Avoid the Mode Collapse

Abstract

Journal

Organization