Auto-Encoding Generative Adversarial Networks towards Mode Collapse Reduction and Feature Representation Enhancement

Yang Zou, Yuxuan Wang, Xiaoxiang Lu · 2023

Read Auto-Encoding Generative Adversarial Networks towards Mode Collapse Reduction and Feature Representation Enhancement by Yang Zou, Yuxuan Wang, Xiaoxiang Lu (2023) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/pmc-10743122.txt.

Description

Yang Zou, Yuxuan Wang, and Xiaoxiang Lu rebuild a picture-making network to fight a known flaw in a 2023 paper. Such networks pit a maker against a judge, but makers often collapse onto a few safe outputs and training shakes. The new design fields a set of makers, plus a coder pair that maps samples into a trait space.

A grouping step watches how real and made samples spread, and a match step keeps the two spreads in line so makers do not crowd one mode. Joint training sharpens the traits the coder pulls out. Tests on two kinds of data sets show fewer collapses and richer traits.

Catalog record

CREATORYang Zou, Yuxuan Wang, Xiaoxiang Lu
YEAR2023
COLLECTIONPAPERS
FORMATTXT
FILE SIZE49 KB
VERSION1.0
LICENSECC BY 4.0
RIGHTSCC BY
SOURCEEurope PMC

Source: https://europepmc.org/articles/PMC10743122

Browse the full library · About the Tower of Babel