Syntax Matters! Syntax-Controlled in Text Style Transfer

TL;DR

This paper highlights the importance of syntax in text style transfer, showing that existing style classifiers are inadequate for syntax learning, and introduces a syntax-aware model that improves transfer quality and fluency.

Contribution

The paper presents a novel syntax-aware style classifier and a new model, SACG, that better captures syntax for improved text style transfer performance.

Findings

SACG outperforms state-of-the-art methods on two TST tasks.

The style classifier's inability to learn syntax limits TST performance.

SACG generates fluent sentences that preserve original content.

Abstract

Existing text style transfer (TST) methods rely on style classifiers to disentangle the text's content and style attributes for text style transfer. While the style classifier plays a critical role in existing TST methods, there is no known investigation on its effect on the TST methods. In this paper, we conduct an empirical study on the limitations of the style classifiers used in existing TST methods. We demonstrate that the existing style classifiers cannot learn sentence syntax effectively and ultimately worsen existing TST models' performance. To address this issue, we propose a novel Syntax-Aware Controllable Generation (SACG) model, which includes a syntax-aware style classifier that ensures learned style latent representations effectively capture the syntax information for TST. Through extensive experiments on two popular TST tasks, we show that our proposed method significantly outperforms the state-of-the-art methods. Our case studies have also demonstrated SACG's ability to generate fluent target-style sentences that preserved the original content.