Improving Semantic Dependency Parsing with Syntactic Features

Robin Kurtz, Daniel Roxbo, Marco Kuhlmann


Abstract
We extend a state-of-the-art deep neural architecture for semantic dependency parsing with features defined over syntactic dependency trees. Our empirical results show that only gold-standard syntactic information leads to consistent improvements in semantic parsing accuracy, and that the magnitude of these improvements varies with the specific combination of the syntactic and the semantic representation used. In contrast, automatically predicted syntax does not seem to help semantic parsing. Our error analysis suggests that there is a significant overlap between syntactic and semantic representations.
Anthology ID:
W19-6202
Volume:
Proceedings of the First NLPL Workshop on Deep Learning for Natural Language Processing
Month:
September
Year:
2019
Address:
Turku, Finland
Venue:
NoDaLiDa
SIG:
Publisher:
Linköping University Electronic Press
Note:
Pages:
12–21
Language:
URL:
https://aclanthology.org/W19-6202
DOI:
Bibkey:
Cite (ACL):
Robin Kurtz, Daniel Roxbo, and Marco Kuhlmann. 2019. Improving Semantic Dependency Parsing with Syntactic Features. In Proceedings of the First NLPL Workshop on Deep Learning for Natural Language Processing, pages 12–21, Turku, Finland. Linköping University Electronic Press.
Cite (Informal):
Improving Semantic Dependency Parsing with Syntactic Features (Kurtz et al., NoDaLiDa 2019)
Copy Citation:
PDF:
https://preview.aclanthology.org/ingestion-script-update/W19-6202.pdf