Demystifying Inductive Biases for (Beta-)VAE Based ArchitecturesDownload PDFOpen Website

Dominik Zietlow, Michal Rolinek, Georg Martius

2021 (modified: 03 Oct 2021)ICML 2021Readers: Everyone
Abstract: The performance of Beta-Variational-Autoencoders and their variants on learning semantically meaningful, disentangled representations is unparalleled. On the other hand, there are theoretical argum...
0 Replies