Depth-Bounded Statistical PCFG Induction as a Model of Human Grammar Acquisition
AbstractThis article describes a simple PCFG induction model with a fixed category domain that predicts a large majority of attested constituent boundaries, and predicts labels consistent with nearly half of attested constituent labels on a standard evaluation data set of child-direc...
| Main Authors: | , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
The MIT Press
2021-03-01
|
| Series: | Computational Linguistics |
| Online Access: | https://direct.mit.edu/coli/article/47/1/181/97336/Depth-Bounded-Statistical-PCFG-Induction-as-a |
| _version_ | 1828219718455001088 |
|---|---|
| author | Lifeng Jin Lane Schwartz Finale Doshi-Velez Timothy Miller William Schuler |
| author_facet | Lifeng Jin Lane Schwartz Finale Doshi-Velez Timothy Miller William Schuler |
| author_sort | Lifeng Jin |
| collection | DOAJ |
| description |
AbstractThis article describes a simple PCFG induction model with a fixed category domain that predicts a large majority of attested constituent boundaries, and predicts labels consistent with nearly half of attested constituent labels on a standard evaluation data set of child-directed speech. The article then explores the idea that the difference between simple grammars exhibited by child learners and fully recursive grammars exhibited by adult learners may be an effect of increasing working memory capacity, where the shallow grammars are constrained images of the recursive grammars. An implementation of these memory bounds as limits on center embedding in a depth-specific transform of a recursive grammar yields a significant improvement over an equivalent but unbounded baseline, suggesting that this arrangement may indeed confer a learning advantage. |
| first_indexed | 2024-04-12T16:19:10Z |
| format | Article |
| id | doaj.art-598ecaefbbee4a0f9ebf9de31c81f4cc |
| institution | Directory Open Access Journal |
| issn | 0891-2017 1530-9312 |
| language | English |
| last_indexed | 2024-04-12T16:19:10Z |
| publishDate | 2021-03-01 |
| publisher | The MIT Press |
| record_format | Article |
| series | Computational Linguistics |
| spelling | doaj.art-598ecaefbbee4a0f9ebf9de31c81f4cc2022-12-22T03:25:37ZengThe MIT PressComputational Linguistics0891-20171530-93122021-03-0147118121610.1162/coli_a_00399Depth-Bounded Statistical PCFG Induction as a Model of Human Grammar AcquisitionLifeng Jin0Lane Schwartz1Finale Doshi-Velez2Timothy Miller3William Schuler4The Ohio State University, Department of Linguistics. jin.544@osu.eduUniversity of Illinois at Urbana-Champaign, Department of Linguistics. lanes@illinois.eduHarvard University, Department of Computer Science. finale@seas.harvard.eduBoston Children’s Hospital & Harvard Medical School, Computational Health Informatics Program. timothy.miller@childrens.harvard.eduThe Ohio State University, Department of Linguistics. schuler@ling.osu.edu AbstractThis article describes a simple PCFG induction model with a fixed category domain that predicts a large majority of attested constituent boundaries, and predicts labels consistent with nearly half of attested constituent labels on a standard evaluation data set of child-directed speech. The article then explores the idea that the difference between simple grammars exhibited by child learners and fully recursive grammars exhibited by adult learners may be an effect of increasing working memory capacity, where the shallow grammars are constrained images of the recursive grammars. An implementation of these memory bounds as limits on center embedding in a depth-specific transform of a recursive grammar yields a significant improvement over an equivalent but unbounded baseline, suggesting that this arrangement may indeed confer a learning advantage.https://direct.mit.edu/coli/article/47/1/181/97336/Depth-Bounded-Statistical-PCFG-Induction-as-a |
| spellingShingle | Lifeng Jin Lane Schwartz Finale Doshi-Velez Timothy Miller William Schuler Depth-Bounded Statistical PCFG Induction as a Model of Human Grammar Acquisition Computational Linguistics |
| title | Depth-Bounded Statistical PCFG Induction as a Model of Human Grammar Acquisition |
| title_full | Depth-Bounded Statistical PCFG Induction as a Model of Human Grammar Acquisition |
| title_fullStr | Depth-Bounded Statistical PCFG Induction as a Model of Human Grammar Acquisition |
| title_full_unstemmed | Depth-Bounded Statistical PCFG Induction as a Model of Human Grammar Acquisition |
| title_short | Depth-Bounded Statistical PCFG Induction as a Model of Human Grammar Acquisition |
| title_sort | depth bounded statistical pcfg induction as a model of human grammar acquisition |
| url | https://direct.mit.edu/coli/article/47/1/181/97336/Depth-Bounded-Statistical-PCFG-Induction-as-a |
| work_keys_str_mv | AT lifengjin depthboundedstatisticalpcfginductionasamodelofhumangrammaracquisition AT laneschwartz depthboundedstatisticalpcfginductionasamodelofhumangrammaracquisition AT finaledoshivelez depthboundedstatisticalpcfginductionasamodelofhumangrammaracquisition AT timothymiller depthboundedstatisticalpcfginductionasamodelofhumangrammaracquisition AT williamschuler depthboundedstatisticalpcfginductionasamodelofhumangrammaracquisition |