Graph-to-Graph Transformer for Transition-based Dependency Parsing

Alireza Mohammadshahi, James Henderson


Abstract
We propose the Graph2Graph Transformer architecture for conditioning on and predicting arbitrary graphs, and apply it to the challenging task of transition-based dependency parsing. After proposing two novel Transformer models of transition-based dependency parsing as strong baselines, we show that adding the proposed mechanisms for conditioning on and predicting graphs of Graph2Graph Transformer results in significant improvements, both with and without BERT pre-training. The novel baselines and their integration with Graph2Graph Transformer significantly outperform the state-of-the-art in traditional transition-based dependency parsing on both English Penn Treebank, and 13 languages of Universal Dependencies Treebanks. Graph2Graph Transformer can be integrated with many previous structured prediction methods, making it easy to apply to a wide range of NLP tasks.
Anthology ID:
2020.findings-emnlp.294
Volume:
Findings of the Association for Computational Linguistics: EMNLP 2020
Month:
November
Year:
2020
Address:
Online
Editors:
Trevor Cohn, Yulan He, Yang Liu
Venue:
Findings
SIG:
Publisher:
Association for Computational Linguistics
Note:
Pages:
3278–3289
Language:
URL:
https://aclanthology.org/2020.findings-emnlp.294
DOI:
10.18653/v1/2020.findings-emnlp.294
Bibkey:
Cite (ACL):
Alireza Mohammadshahi and James Henderson. 2020. Graph-to-Graph Transformer for Transition-based Dependency Parsing. In Findings of the Association for Computational Linguistics: EMNLP 2020, pages 3278–3289, Online. Association for Computational Linguistics.
Cite (Informal):
Graph-to-Graph Transformer for Transition-based Dependency Parsing (Mohammadshahi & Henderson, Findings 2020)
Copy Citation:
PDF:
https://aclanthology.org/2020.findings-emnlp.294.pdf
Optional supplementary material:
 2020.findings-emnlp.294.OptionalSupplementaryMaterial.zip
Video:
 https://slideslive.com/38940653
Code
 alirezamshi/G2GTr
Data
Penn Treebank