@inproceedings{bailly-etal-2023-syntax,
title = "Syntax and Geometry of Information",
author = {Bailly, Rapha{\"e}l and
Leblond, Laurent and
G{\'a}bor, Kata},
editor = "Rogers, Anna and
Boyd-Graber, Jordan and
Okazaki, Naoaki",
booktitle = "Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)",
month = jul,
year = "2023",
address = "Toronto, Canada",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.acl-long.590",
doi = "10.18653/v1/2023.acl-long.590",
pages = "10576--10595",
abstract = "This paper presents an information-theoretical model of syntactic generalization. We study syntactic generalization from the perspective of the capacity to disentangle semantic and structural information, emulating the human capacity to assign a grammaticality judgment to semantically nonsensical sentences. In order to isolate the structure, we propose to represent the probability distribution behind a corpus as the product of the probability of a semantic context and the probability of a structure, the latter being independent of the former. We further elaborate the notion of abstraction as a relaxation of the property of independence. It is based on the measure of structural and contextual information for a given representation. We test abstraction as an optimization objective on the task of inducing syntactic categories from natural language data and show that it significantly outperforms alternative methods. Furthermore, we find that when syntax-unaware optimization objectives succeed in the task, their success is mainly due to an implicit disentanglement process rather than to the model structure. On the other hand, syntactic categories can be deduced in a principled way from the independence between structure and context.",
}
<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="bailly-etal-2023-syntax">
<titleInfo>
<title>Syntax and Geometry of Information</title>
</titleInfo>
<name type="personal">
<namePart type="given">Raphaël</namePart>
<namePart type="family">Bailly</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Laurent</namePart>
<namePart type="family">Leblond</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Kata</namePart>
<namePart type="family">Gábor</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2023-07</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)</title>
</titleInfo>
<name type="personal">
<namePart type="given">Anna</namePart>
<namePart type="family">Rogers</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jordan</namePart>
<namePart type="family">Boyd-Graber</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Naoaki</namePart>
<namePart type="family">Okazaki</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">Toronto, Canada</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
</relatedItem>
<abstract>This paper presents an information-theoretical model of syntactic generalization. We study syntactic generalization from the perspective of the capacity to disentangle semantic and structural information, emulating the human capacity to assign a grammaticality judgment to semantically nonsensical sentences. In order to isolate the structure, we propose to represent the probability distribution behind a corpus as the product of the probability of a semantic context and the probability of a structure, the latter being independent of the former. We further elaborate the notion of abstraction as a relaxation of the property of independence. It is based on the measure of structural and contextual information for a given representation. We test abstraction as an optimization objective on the task of inducing syntactic categories from natural language data and show that it significantly outperforms alternative methods. Furthermore, we find that when syntax-unaware optimization objectives succeed in the task, their success is mainly due to an implicit disentanglement process rather than to the model structure. On the other hand, syntactic categories can be deduced in a principled way from the independence between structure and context.</abstract>
<identifier type="citekey">bailly-etal-2023-syntax</identifier>
<identifier type="doi">10.18653/v1/2023.acl-long.590</identifier>
<location>
<url>https://aclanthology.org/2023.acl-long.590</url>
</location>
<part>
<date>2023-07</date>
<extent unit="page">
<start>10576</start>
<end>10595</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T Syntax and Geometry of Information
%A Bailly, Raphaël
%A Leblond, Laurent
%A Gábor, Kata
%Y Rogers, Anna
%Y Boyd-Graber, Jordan
%Y Okazaki, Naoaki
%S Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)
%D 2023
%8 July
%I Association for Computational Linguistics
%C Toronto, Canada
%F bailly-etal-2023-syntax
%X This paper presents an information-theoretical model of syntactic generalization. We study syntactic generalization from the perspective of the capacity to disentangle semantic and structural information, emulating the human capacity to assign a grammaticality judgment to semantically nonsensical sentences. In order to isolate the structure, we propose to represent the probability distribution behind a corpus as the product of the probability of a semantic context and the probability of a structure, the latter being independent of the former. We further elaborate the notion of abstraction as a relaxation of the property of independence. It is based on the measure of structural and contextual information for a given representation. We test abstraction as an optimization objective on the task of inducing syntactic categories from natural language data and show that it significantly outperforms alternative methods. Furthermore, we find that when syntax-unaware optimization objectives succeed in the task, their success is mainly due to an implicit disentanglement process rather than to the model structure. On the other hand, syntactic categories can be deduced in a principled way from the independence between structure and context.
%R 10.18653/v1/2023.acl-long.590
%U https://aclanthology.org/2023.acl-long.590
%U https://doi.org/10.18653/v1/2023.acl-long.590
%P 10576-10595
Markdown (Informal)
[Syntax and Geometry of Information](https://aclanthology.org/2023.acl-long.590) (Bailly et al., ACL 2023)
ACL
- Raphaël Bailly, Laurent Leblond, and Kata Gábor. 2023. Syntax and Geometry of Information. In Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), pages 10576–10595, Toronto, Canada. Association for Computational Linguistics.