LaDDer: Latent Data Distribution Modelling with a Generative Prior

TL;DR

LaDDer introduces a novel variational autoencoder framework that models the latent data distribution more accurately using a meta-embedding ladder of VAEs, improving representation learning and enabling better latent space interpolation.

Contribution

It proposes LaDDer, a new method employing multiple VAEs and a non-parametric hyper prior to better capture complex latent distributions in variational autoencoders.

Findings

Improves the accuracy of latent distribution estimation

Enhances representation quality in VAEs

Enables effective latent space interpolation

Abstract

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 \textbf{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.