Learned Construction Grammars Converge Across Registers Given Increased Exposure

Type of content
Conference Contributions - Published
Publisher's DOI/URI
Thesis discipline
Degree name
Publisher
Association for Computational Linguistics
Journal Title
Journal ISSN
Volume Title
Language
Date
2021
Authors
Tayyar Madabushi H
Dunn, Jonathan
Abstract

This paper measures the impact of increased exposure on whether learned construction grammars converge onto shared representations when trained on data from different registers. Register influences the frequency of constructions, with some structures common in formal but not informal usage. We expect that a grammar induction algorithm exposed to different registers will acquire different constructions. To what degree does increased exposure lead to the convergence of register-specific grammars? The experiments in this paper simulate language learning in 12 languages (half Germanic and half Romance) with corpora representing three registers (Twitter, Wikipedia, Web). These simulations are repeated with increasing amounts of exposure, from 100k to 2 million words, to measure the impact of exposure on the convergence of grammars. The results show that increased exposure does lead to converging grammars across all languages. In addition, a shared core of register-universal constructions remains constant across increasing amounts of exposure.

Description
Citation
Dunn J, Tayyar Madabushi H (2021). Learned Construction Grammars Converge Across Registers Given Increased Exposure. Conference on Natural Language Learning (CoNLL). Proceedings of the Conference on Natural Language Learning.
Keywords
Ngā upoko tukutuku/Māori subject headings
ANZSRC fields of research
Fields of Research::47 - Language, communication and culture::4704 - Linguistics::470403 - Computational linguistics
Fields of Research::47 - Language, communication and culture::4704 - Linguistics::470409 - Linguistic structures (incl. phonology, morphology and syntax)
Fields of Research::47 - Language, communication and culture::4704 - Linguistics::470404 - Corpus linguistics
Fields of Research::47 - Language, communication and culture::4704 - Linguistics::470401 - Applied linguistics and educational linguistics
Fields of Research::47 - Language, communication and culture::4703 - Language studies::470304 - Comparative language studies
Rights
All rights reserved unless otherwise stated