Enhancing Content Preservation in Text Style Transfer Using Reverse Attention and Conditional Layer Normalization

Dongkyu Lee, Zhiliang Tian, Lanqing Xue, Nevin L. Zhang


Abstract
Text style transfer aims to alter the style (e.g., sentiment) of a sentence while preserving its content. A common approach is to map a given sentence to content representation that is free of style, and the content representation is fed to a decoder with a target style. Previous methods in filtering style completely remove tokens with style at the token level, which incurs the loss of content information. In this paper, we propose to enhance content preservation by implicitly removing the style information of each token with reverse attention, and thereby retain the content. Furthermore, we fuse content information when building the target style representation, making it dynamic with respect to the content. Our method creates not only style-independent content representation, but also content-dependent style representation in transferring style. Empirical results show that our method outperforms the state-of-the-art baselines by a large margin in terms of content preservation. In addition, it is also competitive in terms of style transfer accuracy and fluency.
Anthology ID:
2021.acl-long.8
Volume:
Proceedings of the 59th Annual Meeting of the Association for Computational Linguistics and the 11th International Joint Conference on Natural Language Processing (Volume 1: Long Papers)
Month:
August
Year:
2021
Address:
Online
Venues:
ACL | IJCNLP
SIG:
Publisher:
Association for Computational Linguistics
Note:
Pages:
93–102
Language:
URL:
https://aclanthology.org/2021.acl-long.8
DOI:
10.18653/v1/2021.acl-long.8
Bibkey:
Cite (ACL):
Dongkyu Lee, Zhiliang Tian, Lanqing Xue, and Nevin L. Zhang. 2021. Enhancing Content Preservation in Text Style Transfer Using Reverse Attention and Conditional Layer Normalization. In Proceedings of the 59th Annual Meeting of the Association for Computational Linguistics and the 11th International Joint Conference on Natural Language Processing (Volume 1: Long Papers), pages 93–102, Online. Association for Computational Linguistics.
Cite (Informal):
Enhancing Content Preservation in Text Style Transfer Using Reverse Attention and Conditional Layer Normalization (Lee et al., ACL-IJCNLP 2021)
Copy Citation:
PDF:
https://aclanthology.org/2021.acl-long.8.pdf
Video:
 https://aclanthology.org/2021.acl-long.8.mp4
Code
 MovingKyu/RACoLN