Leveraging Plug-and-Play Models for Rhetorical Structure Control in Text Generation

Yuka Yokogawa, Tatsuya Ishigaki, Hiroya Takamura, Yusuke Miyao, Ichiro Kobayashi


Abstract
We propose a method that extends a BART-based language generator using a plug-and-play model to control the rhetorical structure of generated text. Our approach considers rhetorical relations between clauses and generates sentences that reflect this structure using plug-and-play language models. We evaluated our method using the Newsela corpus, which consists of texts at various levels of English proficiency. Our experiments demonstrated that our method outperforms the vanilla BART in terms of the correctness of output discourse and rhetorical structures. In existing methods, the rhetorical structure tends to deteriorate when compared to the baseline, the vanilla BART, as measured by n-gram overlap metrics such as BLEU. However, our proposed method does not exhibit this significant deterioration, demonstrating its advantage.
Anthology ID:
2024.inlg-main.40
Volume:
Proceedings of the 17th International Natural Language Generation Conference
Month:
September
Year:
2024
Address:
Tokyo, Japan
Editors:
Saad Mahamood, Nguyen Le Minh, Daphne Ippolito
Venue:
INLG
SIG:
SIGGEN
Publisher:
Association for Computational Linguistics
Note:
Pages:
486–493
Language:
URL:
https://aclanthology.org/2024.inlg-main.40
DOI:
Bibkey:
Cite (ACL):
Yuka Yokogawa, Tatsuya Ishigaki, Hiroya Takamura, Yusuke Miyao, and Ichiro Kobayashi. 2024. Leveraging Plug-and-Play Models for Rhetorical Structure Control in Text Generation. In Proceedings of the 17th International Natural Language Generation Conference, pages 486–493, Tokyo, Japan. Association for Computational Linguistics.
Cite (Informal):
Leveraging Plug-and-Play Models for Rhetorical Structure Control in Text Generation (Yokogawa et al., INLG 2024)
Copy Citation:
PDF:
https://aclanthology.org/2024.inlg-main.40.pdf
Supplementary attachment:
 2024.inlg-main.40.Supplementary_Attachment.pdf