LaDDer: Latent Data Distribution Modelling with a Generative Prior

08/31/2020
by   Shuyu Lin, et al.
14

In this paper, we show that the performance of a learnt generative model is closely related to the model's ability to accurately represent the inferred latent data distribution, i.e. its topology and structural properties. We propose LaDDer to achieve accurate modelling of the latent data distribution in a variational autoencoder framework and to facilitate better representation learning. The central idea of LaDDer is a meta-embedding concept, which uses multiple VAE models to learn an embedding of the embeddings, forming a ladder of encodings. We use a non-parametric mixture as the hyper prior for the innermost VAE and learn all the parameters in a unified variational framework. From extensive experiments, we show that our LaDDer model is able to accurately estimate complex latent distribution and results in improvement in the representation quality. We also propose a novel latent space interpolation method that utilises the derived data distribution.

READ FULL TEXT

Please sign up or login with your details

Forgot password? Click here to reset