Domain Adaptive Text Style Transfer

Abstract

Text style transfer without parallel data has achieved some practical success. However, in the scenario where less data is available, these methods may yield poor performance. In this paper, we examine domain adaptation for text style transfer to leverage massively available data from other domains. These data may demonstrate domain shift, which impedes the benefits of utilizing such data for training. To address this challenge, we propose simple yet effective domain adaptive text style transfer models, enabling domain-adaptive information exchange. The proposed models presumably learn from the source domain to: (i) distinguish stylized information and generic content information; (ii) maximally preserve content information; and (iii) adaptively transfer the styles in a domain-aware manner. We evaluate the proposed models on two style transfer tasks (sentiment and formality) over multiple target domains where only limited non-parallel data is available. Extensive experiments demonstrate the effectiveness of the proposed model compared to the baselines.

arXiv commentEMNLP 2019, long paper

References in corpus (6)

• An Empirical Investigation of Catastrophic Forgetting in Gradient-Based Neural Networks
• Style Transfer from Non-Parallel Text by Cross-Alignment
• Personalizing Dialogue Agents: I have a dog, do you have pets too?
• Adversarial Ranking for Language Generation
• Content preserving text generation with attribute controls
• Multiple-Attribute Text Style Transfer

Cited by in corpus (4)

• From Theories on Styles to their Transfer in Text: Bridging the Gap with a Hierarchical Survey
• What Makes A Good Story? Designing Composite Rewards for Visual Storytelling
• Plug and Play Autoencoders for Conditional Text Generation
• Learning to Generate Multiple Style Transfer Outputs for an Input Sentence