Recursive Subtree Composition in LSTM-Based Dependency Parsing

Miryam de Lhoneux, Miguel Ballesteros, Joakim Nivre


Abstract
The need for tree structure modelling on top of sequence modelling is an open issue in neural dependency parsing. We investigate the impact of adding a tree layer on top of a sequential model by recursively composing subtree representations (composition) in a transition-based parser that uses features extracted by a BiLSTM. Composition seems superfluous with such a model, suggesting that BiLSTMs capture information about subtrees. We perform model ablations to tease out the conditions under which composition helps. When ablating the backward LSTM, performance drops and composition does not recover much of the gap. When ablating the forward LSTM, performance drops less dramatically and composition recovers a substantial part of the gap, indicating that a forward LSTM and composition capture similar information. We take the backward LSTM to be related to lookahead features and the forward LSTM to the rich history-based features both crucial for transition-based parsers. To capture history-based information, composition is better than a forward LSTM on its own, but it is even better to have a forward LSTM as part of a BiLSTM. We correlate results with language properties, showing that the improved lookahead of a backward LSTM is especially important for head-final languages.
Anthology ID:
N19-1159
Original:
N19-1159v1
Version 2:
N19-1159v2
Volume:
Proceedings of the 2019 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, Volume 1 (Long and Short Papers)
Month:
June
Year:
2019
Address:
Minneapolis, Minnesota
Editors:
Jill Burstein, Christy Doran, Thamar Solorio
Venue:
NAACL
SIG:
Publisher:
Association for Computational Linguistics
Note:
Pages:
1566–1576
Language:
URL:
https://aclanthology.org/N19-1159
DOI:
10.18653/v1/N19-1159
Bibkey:
Cite (ACL):
Miryam de Lhoneux, Miguel Ballesteros, and Joakim Nivre. 2019. Recursive Subtree Composition in LSTM-Based Dependency Parsing. In Proceedings of the 2019 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, Volume 1 (Long and Short Papers), pages 1566–1576, Minneapolis, Minnesota. Association for Computational Linguistics.
Cite (Informal):
Recursive Subtree Composition in LSTM-Based Dependency Parsing (de Lhoneux et al., NAACL 2019)
Copy Citation:
PDF:
https://aclanthology.org/N19-1159.pdf
Presentation:
 N19-1159.Presentation.pdf
Video:
 https://aclanthology.org/N19-1159.mp4
Code
 mdelhoneux/uuparser-composition