On Disentanglement and Mutual Information in Semi-Supervised Variational Auto-Encoders

Elliott Gordon Rodriguez; Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) Workshops, 2021, pp. 1257-1262

Abstract


In the context of variational auto-encoders, learning disentangled latent variable representations remains a challenging problem. In this abstract, we consider the semi-supervised setting, in which the factors of variation are labelled for a small fraction of our samples. We examine how the quality of learned representations is affected by the dimension of the unsupervised component of the latent space. We also consider a variational lower bound for the mutual information between the data and the semi-supervised component of the latent space, and analyze its role in the context of disentangled representation learning.

Related Material


[pdf]
[bibtex]
@InProceedings{Rodriguez_2021_CVPR, author = {Rodriguez, Elliott Gordon}, title = {On Disentanglement and Mutual Information in Semi-Supervised Variational Auto-Encoders}, booktitle = {Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) Workshops}, month = {June}, year = {2021}, pages = {1257-1262} }