Semantic Dependency Parsing with Edge GNNs

Songlin Yang, Kewei Tu


Abstract
Second-order neural parsers have obtained high accuracy in semantic dependency parsing. Inspired by the factor graph representation of second-order parsing, we propose edge graph neural networks (E-GNNs). In an E-GNN, each node corresponds to a dependency edge, and the neighbors are defined in terms of sibling, co-parent, and grandparent relationships. We conduct experiments on SemEval 2015 Task 18 English datasets, showing the superior performance of E-GNNs.
Anthology ID:
2022.findings-emnlp.452
Volume:
Findings of the Association for Computational Linguistics: EMNLP 2022
Month:
December
Year:
2022
Address:
Abu Dhabi, United Arab Emirates
Editors:
Yoav Goldberg, Zornitsa Kozareva, Yue Zhang
Venue:
Findings
SIG:
Publisher:
Association for Computational Linguistics
Note:
Pages:
6096–6102
Language:
URL:
https://aclanthology.org/2022.findings-emnlp.452
DOI:
10.18653/v1/2022.findings-emnlp.452
Bibkey:
Cite (ACL):
Songlin Yang and Kewei Tu. 2022. Semantic Dependency Parsing with Edge GNNs. In Findings of the Association for Computational Linguistics: EMNLP 2022, pages 6096–6102, Abu Dhabi, United Arab Emirates. Association for Computational Linguistics.
Cite (Informal):
Semantic Dependency Parsing with Edge GNNs (Yang & Tu, Findings 2022)
Copy Citation:
PDF:
https://aclanthology.org/2022.findings-emnlp.452.pdf