Discontinuous Combinatory Constituency Parsing

AbstractWe extend a pair of continuous combinator-based constituency parsers (one binary and one multi-branching) into a discontinuous pair. Our parsers iteratively compose constituent vectors from word embeddings without any grammar constraints. Their empirical complexities are subq...

Full description

Bibliographic Details
Main Authors: Zhousi Chen, Mamoru Komachi
Format: Article
Language:English
Published: The MIT Press 2023-01-01
Series:Transactions of the Association for Computational Linguistics
Online Access:https://direct.mit.edu/tacl/article/doi/10.1162/tacl_a_00546/115348/Discontinuous-Combinatory-Constituency-Parsing
_version_ 1797796574307287040
author Zhousi Chen
Mamoru Komachi
author_facet Zhousi Chen
Mamoru Komachi
author_sort Zhousi Chen
collection DOAJ
description AbstractWe extend a pair of continuous combinator-based constituency parsers (one binary and one multi-branching) into a discontinuous pair. Our parsers iteratively compose constituent vectors from word embeddings without any grammar constraints. Their empirical complexities are subquadratic. Our extension includes 1) a swap action for the orientation-based binary model and 2) biaffine attention for the chunker-based multi-branching model. In tests conducted with the Discontinuous Penn Treebank and TIGER Treebank, we achieved state-of-the-art discontinuous accuracy with a significant speed advantage.
first_indexed 2024-03-13T03:36:05Z
format Article
id doaj.art-13c5b506bfb14753baabb2ac9c8ed0fb
institution Directory Open Access Journal
issn 2307-387X
language English
last_indexed 2024-03-13T03:36:05Z
publishDate 2023-01-01
publisher The MIT Press
record_format Article
series Transactions of the Association for Computational Linguistics
spelling doaj.art-13c5b506bfb14753baabb2ac9c8ed0fb2023-06-23T18:59:01ZengThe MIT PressTransactions of the Association for Computational Linguistics2307-387X2023-01-011126728310.1162/tacl_a_00546Discontinuous Combinatory Constituency ParsingZhousi Chen0Mamoru Komachi1Faculty of Systems Design Tokyo Metropolitan University, 6-6 Asahigaoka, Hino, Tokyo 191-0065, Japan. chen-zhousi@ed.tmu.ac.jpFaculty of Systems Design Tokyo Metropolitan University, 6-6 Asahigaoka, Hino, Tokyo 191-0065, Japan. komachi@tmu.ac.jp AbstractWe extend a pair of continuous combinator-based constituency parsers (one binary and one multi-branching) into a discontinuous pair. Our parsers iteratively compose constituent vectors from word embeddings without any grammar constraints. Their empirical complexities are subquadratic. Our extension includes 1) a swap action for the orientation-based binary model and 2) biaffine attention for the chunker-based multi-branching model. In tests conducted with the Discontinuous Penn Treebank and TIGER Treebank, we achieved state-of-the-art discontinuous accuracy with a significant speed advantage.https://direct.mit.edu/tacl/article/doi/10.1162/tacl_a_00546/115348/Discontinuous-Combinatory-Constituency-Parsing
spellingShingle Zhousi Chen
Mamoru Komachi
Discontinuous Combinatory Constituency Parsing
Transactions of the Association for Computational Linguistics
title Discontinuous Combinatory Constituency Parsing
title_full Discontinuous Combinatory Constituency Parsing
title_fullStr Discontinuous Combinatory Constituency Parsing
title_full_unstemmed Discontinuous Combinatory Constituency Parsing
title_short Discontinuous Combinatory Constituency Parsing
title_sort discontinuous combinatory constituency parsing
url https://direct.mit.edu/tacl/article/doi/10.1162/tacl_a_00546/115348/Discontinuous-Combinatory-Constituency-Parsing
work_keys_str_mv AT zhousichen discontinuouscombinatoryconstituencyparsing
AT mamorukomachi discontinuouscombinatoryconstituencyparsing