Linear Neural Parsing and Hybrid Enhancement for Enhanced Universal Dependencies

Giuseppe Attardi, Daniele Sartiano, Maria Simi


Abstract
To accomplish the shared task on dependency parsing we explore the use of a linear transition-based neural dependency parser as well as a combination of three of them by means of a linear tree combination algorithm. We train separate models for each language on the shared task data. We compare our base parser with two biaffine parsers and also present an ensemble combination of all five parsers, which achieves an average UAS 1.88 point lower than the top official submission. For producing the enhanced dependencies, we exploit a hybrid approach, coupling an algorithmic graph transformation of the dependency tree with predictions made by a multitask machine learning model.
Anthology ID:
2020.iwpt-1.21
Volume:
Proceedings of the 16th International Conference on Parsing Technologies and the IWPT 2020 Shared Task on Parsing into Enhanced Universal Dependencies
Month:
July
Year:
2020
Address:
Online
Venues:
ACL | IWPT | WS
SIG:
SIGPARSE
Publisher:
Association for Computational Linguistics
Note:
Pages:
206–214
Language:
URL:
https://aclanthology.org/2020.iwpt-1.21
DOI:
10.18653/v1/2020.iwpt-1.21
Bibkey:
Cite (ACL):
Giuseppe Attardi, Daniele Sartiano, and Maria Simi. 2020. Linear Neural Parsing and Hybrid Enhancement for Enhanced Universal Dependencies. In Proceedings of the 16th International Conference on Parsing Technologies and the IWPT 2020 Shared Task on Parsing into Enhanced Universal Dependencies, pages 206–214, Online. Association for Computational Linguistics.
Cite (Informal):
Linear Neural Parsing and Hybrid Enhancement for Enhanced Universal Dependencies (Attardi et al., IWPT 2020)
Copy Citation:
PDF:
https://aclanthology.org/2020.iwpt-1.21.pdf
Video:
 http://slideslive.com/38929688