Generating Sentences from a Continuous Space

TL;DR

This paper introduces a variational autoencoder model for sentences that captures global sentence properties in a continuous latent space, enabling diverse sentence generation and interpolation.

Contribution

It presents a novel RNN-based variational autoencoder that explicitly models holistic sentence features and explores its applications and limitations.

Findings

Generated diverse, well-formed sentences from the prior

Enabled interpolation between sentences in latent space

Showed effectiveness in imputing missing words

Abstract

The standard recurrent neural network language model (RNNLM) generates sentences one word at a time and does not work from an explicit global sentence representation. In this work, we introduce and study an RNN-based variational autoencoder generative model that incorporates distributed latent representations of entire sentences. This factorization allows it to explicitly model holistic properties of sentences such as style, topic, and high-level syntactic features. Samples from the prior over these sentence representations remarkably produce diverse and well-formed sentences through simple deterministic decoding. By examining paths through this latent space, we are able to generate coherent novel sentences that interpolate between known sentences. We present techniques for solving the difficult learning problem presented by this model, demonstrate its effectiveness in imputing missing words, explore many interesting properties of the model's latent sentence space, and present negative results on the use of the model in language modeling.