X-Ray Sobolev Variational Auto-Encoders
share
The quality of the generative models (Generative adversarial networks, Variational Auto-Encoders, ...) depends heavily on the choice of a good probability distance. However some popular metrics lack convenient properties such as (geodesic) convexity, fast evaluation and so on. To address these shortcomings, we introduce a class of distances that have built-in convexity. We investigate the relationship with some known paradigms (sliced distances, reproducing kernel Hilbert spaces, energy distances). The distances are shown to posses fast implementations and are included in an adapted Variational Auto-Encoder termed X-ray Sobolev Variational Auto-Encoder (XS-VAE) which produces good quality results on standard generative datasets.
READ FULL TEXT
