We gratefully acknowledge support from
the Simons Foundation and member institutions.
Full-text links:

Download:

Current browse context:

cs

Change to browse by:

References & Citations

DBLP - CS Bibliography

Bookmark

(what is this?)
CiteULike logo BibSonomy logo Mendeley logo del.icio.us logo Digg logo Reddit logo ScienceWISE logo

Computer Science > Machine Learning

Title: Fostering Diversity in Spatial Evolutionary Generative Adversarial Networks

Abstract: Generative adversary networks (GANs) suffer from training pathologies such as instability and mode collapse, which mainly arise from a lack of diversity in their adversarial interactions. Co-evolutionary GAN (CoE-GAN) training algorithms have shown to be resilient to these pathologies. This article introduces Mustangs, a spatially distributed CoE-GAN, which fosters diversity by using different loss functions during the training. Experimental analysis on MNIST and CelebA demonstrated that Mustangs trains statistically more accurate generators.
Comments: Accepted to be presented during Conference of the Spanish Association of Artificial Intelligence (CAEPIA 2021). arXiv admin note: substantial text overlap with arXiv:1905.12702
Subjects: Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Distributed, Parallel, and Cluster Computing (cs.DC)
Cite as: arXiv:2106.13590 [cs.LG]
  (or arXiv:2106.13590v1 [cs.LG] for this version)

Submission history

From: Jamal Toutouh [view email]
[v1] Fri, 25 Jun 2021 12:40:36 GMT (3399kb)

Link back to: arXiv, form interface, contact.