Robust Variational Autoencoder

Haleh Akrami, Anand A. Joshi, Jian Li +2

Machine learning methods often need a large amount of labeled training data. Since the training data is assumed to be the ground truth, outliers can severely degrade learned repres…