Variational Memory Addressing in Generative Models

TL;DR

This paper introduces a variational approach to memory addressing in generative models, enabling effective training and improved multimodal generation by treating memory addresses as latent variables within a probabilistic framework.

Contribution

It presents a novel variational inference method for stochastic memory addressing in generative models, enhancing multimodal data modeling and interpretability of memory contributions.

Findings

Successfully incorporated into a variational autoencoder for few-shot learning

Effectively identifies relevant memory contents with hundreds of unseen characters

Improves multimodal generation and interpretability of memory contributions

Abstract

Aiming to augment generative models with external memory, we interpret the output of a memory module with stochastic addressing as a conditional mixture distribution, where a read operation corresponds to sampling a discrete memory address and retrieving the corresponding content from memory. This perspective allows us to apply variational inference to memory addressing, which enables effective training of the memory module by using the target information to guide memory lookups. Stochastic addressing is particularly well-suited for generative models as it naturally encourages multimodality which is a prominent aspect of most high-dimensional datasets. Treating the chosen address as a latent variable also allows us to quantify the amount of information gained with a memory lookup and measure the contribution of the memory module to the generative process. To illustrate the advantages of this approach we incorporate it into a variational autoencoder and apply the resulting model to the task of generative few-shot learning. The intuition behind this architecture is that the memory module can pick a relevant template from memory and the continuous part of the model can concentrate on modeling remaining variations. We demonstrate empirically that our model is able to identify and access the relevant memory contents even with hundreds of unseen Omniglot characters in memory