Arbitrarily-conditioned Data Imputation

Publication

M. Carvalho, T. Durand, J. He, N. Mehrasa, G. Mori

2nd Symposium on Advances in Approximate Bayesian Inference (AABI), 2019

In this paper, we propose an arbitrarily-conditioned data imputation framework built upon variational autoencoders and normalizing flows. The proposed model is capable of mapping any partial data to a multi-modal latent variational distribution. Sampling from such a distribution leads to stochastic imputation. Preliminary evaluation on MNIST dataset shows promising stochastic imputation conditioned on partial images as input.

Bibtex

@InProceedings{Carvahlo2019,
title ={{Arbitrarily-conditioned Data Imputation}},
author ={Carvalho, Micael and Durand, Thibaut and He, Jiawei and Mehrasa, Nazanin and Mori, Greg},
booktitle ={Proceedings of The 2nd Symposium on Advances in Approximate Bayesian Inference},
year ={2019},
url = {https://openreview.net/forum?id=r1eP5khVKB},

Related Research