Arbitrarily-conditioned Data Imputation
2019-10-16pproximateinference AABI Symposium 2019Unverified0· sign in to hype
Micael Carvalho, Thibaut Durand, JiaWei He, Nazanin Mehrasa, Greg Mori
Unverified — Be the first to reproduce this paper.
ReproduceAbstract
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.