Revisiting the Binary Linearization Technique for Surface Realization

Yevgeniy Puzikov, Claire Gardent, Ido Dagan, Iryna Gurevych


Abstract
End-to-end neural approaches have achieved state-of-the-art performance in many natural language processing (NLP) tasks. Yet, they often lack transparency of the underlying decision-making process, hindering error analysis and certain model improvements. In this work, we revisit the binary linearization approach to surface realization, which exhibits more interpretable behavior, but was falling short in terms of prediction accuracy. We show how enriching the training data to better capture word order constraints almost doubles the performance of the system. We further demonstrate that encoding both local and global prediction contexts yields another considerable performance boost. With the proposed modifications, the system which ranked low in the latest shared task on multilingual surface realization now achieves best results in five out of ten languages, while being on par with the state-of-the-art approaches in others.
Anthology ID:
W19-8635
Volume:
Proceedings of the 12th International Conference on Natural Language Generation
Month:
October–November
Year:
2019
Address:
Tokyo, Japan
Editors:
Kees van Deemter, Chenghua Lin, Hiroya Takamura
Venue:
INLG
SIG:
SIGGEN
Publisher:
Association for Computational Linguistics
Note:
Pages:
268–278
Language:
URL:
https://aclanthology.org/W19-8635
DOI:
10.18653/v1/W19-8635
Bibkey:
Cite (ACL):
Yevgeniy Puzikov, Claire Gardent, Ido Dagan, and Iryna Gurevych. 2019. Revisiting the Binary Linearization Technique for Surface Realization. In Proceedings of the 12th International Conference on Natural Language Generation, pages 268–278, Tokyo, Japan. Association for Computational Linguistics.
Cite (Informal):
Revisiting the Binary Linearization Technique for Surface Realization (Puzikov et al., INLG 2019)
Copy Citation:
PDF:
https://preview.aclanthology.org/naacl-24-ws-corrections/W19-8635.pdf