Unsupervised Text Style Transfer with Deep Generative Models

TL;DR

This paper introduces a versatile unsupervised text style transfer framework using deep generative models, which models content and style as latent codes and unifies previous methods, achieving competitive results.

Contribution

The authors propose a novel framework that models content and style as latent variables, unifies existing approaches, and offers a new perspective on unsupervised text style transfer.

Findings

Achieves better or competitive results on three benchmarks.

Unifies previous embedding and prototype methods.

Provides a principled explanation for existing techniques.

Abstract

We present a general framework for unsupervised text style transfer with deep generative models. The framework models each sentence-label pair in the non-parallel corpus as partially observed from a complete quadruplet which additionally contains two latent codes representing the content and style, respectively. These codes are learned by exploiting dependencies inside the observed data. Then a sentence is transferred by manipulating them. Our framework is able to unify previous embedding and prototype methods as two special forms. It also provides a principled perspective to explain previously proposed techniques in the field such as aligned encoder and adversarial training. We further conduct experiments on three benchmarks. Both automatic and human evaluation results show that our methods achieve better or competitive results compared to several strong baselines.