From EM-Projections to Variational Auto-Encoder

Abstract

This paper reviews the em-projections in information geometry and the recent understanding of variational auto-encoder, and explains that they share a common formulation as joint minimization of the Kullback-Leibler divergence between two manifolds of probability distributions, and the joint minimization can be implemented by alternating projections or alternating gradient descent.

PDF NeurIPSW OpenReview Semantic Scholar

Cite

Text

Han et al. "From EM-Projections to Variational Auto-Encoder." NeurIPS 2020 Workshops: DL-IG, 2020.

Markdown

[Han et al. "From EM-Projections to Variational Auto-Encoder." NeurIPS 2020 Workshops: DL-IG, 2020.](https://mlanthology.org/neuripsw/2020/han2020neuripsw-emprojections/)

BibTeX

@inproceedings{han2020neuripsw-emprojections,
  title     = {{From EM-Projections to Variational Auto-Encoder}},
  author    = {Han, Tian and Zhang, Jun and Wu, Ying Nian},
  booktitle = {NeurIPS 2020 Workshops: DL-IG},
  year      = {2020},
  url       = {https://mlanthology.org/neuripsw/2020/han2020neuripsw-emprojections/}
}