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

CVPRW 2021 pp. 1257-1262

doi:10.1109/CVPRW53098.2021.00138 /cvprw/2021/gordonrodriguez2021cvprw-disentanglement/

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.

CVPRW Semantic Scholar

Cite

Text

Gordon-Rodríguez. "On Disentanglement and Mutual Information in Semi-Supervised Variational Auto-Encoders." IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops, 2021. doi:10.1109/CVPRW53098.2021.00138

Markdown

[Gordon-Rodríguez. "On Disentanglement and Mutual Information in Semi-Supervised Variational Auto-Encoders." IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops, 2021.](https://mlanthology.org/cvprw/2021/gordonrodriguez2021cvprw-disentanglement/) doi:10.1109/CVPRW53098.2021.00138

BibTeX

@inproceedings{gordonrodriguez2021cvprw-disentanglement,
  title     = {{On Disentanglement and Mutual Information in Semi-Supervised Variational Auto-Encoders}},
  author    = {Gordon-Rodríguez, Elliott},
  booktitle = {IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops},
  year      = {2021},
  pages     = {1257-1262},
  doi       = {10.1109/CVPRW53098.2021.00138},
  url       = {https://mlanthology.org/cvprw/2021/gordonrodriguez2021cvprw-disentanglement/}
}