Style Transfer in Text: Exploration and Evaluation

TL;DR

This paper introduces models for text style transfer using non-parallel data, leveraging adversarial networks to separate content and style, and proposes new evaluation metrics that align well with human judgments.

Contribution

The paper presents novel models for style transfer without parallel data and introduces evaluation metrics that effectively measure transfer strength and content preservation.

Findings

Models achieve higher style transfer strength than auto-encoders.

Content preservation metric correlates highly with human judgments.

Proposed models perform well on paper-news and review transfer tasks.

Abstract

Style transfer is an important problem in natural language processing (NLP). However, the progress in language style transfer is lagged behind other domains, such as computer vision, mainly because of the lack of parallel data and principle evaluation metrics. In this paper, we propose to learn style transfer with non-parallel data. We explore two models to achieve this goal, and the key idea behind the proposed models is to learn separate content representations and style representations using adversarial networks. We also propose novel evaluation metrics which measure two aspects of style transfer: transfer strength and content preservation. We access our models and the evaluation metrics on two tasks: paper-news title transfer, and positive-negative review transfer. Results show that the proposed content preservation metric is highly correlate to human judgments, and the proposed models are able to generate sentences with higher style transfer strength and similar content preservation score comparing to auto-encoder.