@inproceedings{park-kim-2024-structural,
title = "Structural Optimization Ambiguity and Simplicity Bias in Unsupervised Neural Grammar Induction",
author = "Park, Jinwook and
Kim, Kangil",
editor = "Ku, Lun-Wei and
Martins, Andre and
Srikumar, Vivek",
booktitle = "Findings of the Association for Computational Linguistics ACL 2024",
month = aug,
year = "2024",
address = "Bangkok, Thailand and virtual meeting",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2024.findings-acl.898",
pages = "15124--15139",
abstract = "Neural parameterization has significantly advanced unsupervised grammar induction. However, training these models with a traditional likelihood loss for all possible parses exacerbates two issues: 1) *structural optimization ambiguity* that arbitrarily selects one among structurally ambiguous optimal grammars despite the specific preference of gold parses, and 2) *structural simplicity bias* that leads a model to underutilize rules to compose parse trees. These challenges subject unsupervised neural grammar induction (UNGI) to inevitable prediction errors, high variance, and the necessity for extensive grammars to achieve accurate predictions. This paper tackles these issues, offering a comprehensive analysis of their origins. As a solution, we introduce *sentence-wise parse-focusing* to reduce the parse pool per sentence for loss evaluation, using the structural bias from pre-trained parsers on the same dataset.In unsupervised parsing benchmark tests, our method significantly improves performance while effectively reducing variance and bias toward overly simplistic parses. Our research promotes learning more compact, accurate, and consistent explicit grammars, facilitating better interpretability.",
}
<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="park-kim-2024-structural">
<titleInfo>
<title>Structural Optimization Ambiguity and Simplicity Bias in Unsupervised Neural Grammar Induction</title>
</titleInfo>
<name type="personal">
<namePart type="given">Jinwook</namePart>
<namePart type="family">Park</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Kangil</namePart>
<namePart type="family">Kim</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2024-08</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Findings of the Association for Computational Linguistics ACL 2024</title>
</titleInfo>
<name type="personal">
<namePart type="given">Lun-Wei</namePart>
<namePart type="family">Ku</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Andre</namePart>
<namePart type="family">Martins</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Vivek</namePart>
<namePart type="family">Srikumar</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">Bangkok, Thailand and virtual meeting</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
</relatedItem>
<abstract>Neural parameterization has significantly advanced unsupervised grammar induction. However, training these models with a traditional likelihood loss for all possible parses exacerbates two issues: 1) *structural optimization ambiguity* that arbitrarily selects one among structurally ambiguous optimal grammars despite the specific preference of gold parses, and 2) *structural simplicity bias* that leads a model to underutilize rules to compose parse trees. These challenges subject unsupervised neural grammar induction (UNGI) to inevitable prediction errors, high variance, and the necessity for extensive grammars to achieve accurate predictions. This paper tackles these issues, offering a comprehensive analysis of their origins. As a solution, we introduce *sentence-wise parse-focusing* to reduce the parse pool per sentence for loss evaluation, using the structural bias from pre-trained parsers on the same dataset.In unsupervised parsing benchmark tests, our method significantly improves performance while effectively reducing variance and bias toward overly simplistic parses. Our research promotes learning more compact, accurate, and consistent explicit grammars, facilitating better interpretability.</abstract>
<identifier type="citekey">park-kim-2024-structural</identifier>
<location>
<url>https://aclanthology.org/2024.findings-acl.898</url>
</location>
<part>
<date>2024-08</date>
<extent unit="page">
<start>15124</start>
<end>15139</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T Structural Optimization Ambiguity and Simplicity Bias in Unsupervised Neural Grammar Induction
%A Park, Jinwook
%A Kim, Kangil
%Y Ku, Lun-Wei
%Y Martins, Andre
%Y Srikumar, Vivek
%S Findings of the Association for Computational Linguistics ACL 2024
%D 2024
%8 August
%I Association for Computational Linguistics
%C Bangkok, Thailand and virtual meeting
%F park-kim-2024-structural
%X Neural parameterization has significantly advanced unsupervised grammar induction. However, training these models with a traditional likelihood loss for all possible parses exacerbates two issues: 1) *structural optimization ambiguity* that arbitrarily selects one among structurally ambiguous optimal grammars despite the specific preference of gold parses, and 2) *structural simplicity bias* that leads a model to underutilize rules to compose parse trees. These challenges subject unsupervised neural grammar induction (UNGI) to inevitable prediction errors, high variance, and the necessity for extensive grammars to achieve accurate predictions. This paper tackles these issues, offering a comprehensive analysis of their origins. As a solution, we introduce *sentence-wise parse-focusing* to reduce the parse pool per sentence for loss evaluation, using the structural bias from pre-trained parsers on the same dataset.In unsupervised parsing benchmark tests, our method significantly improves performance while effectively reducing variance and bias toward overly simplistic parses. Our research promotes learning more compact, accurate, and consistent explicit grammars, facilitating better interpretability.
%U https://aclanthology.org/2024.findings-acl.898
%P 15124-15139
Markdown (Informal)
[Structural Optimization Ambiguity and Simplicity Bias in Unsupervised Neural Grammar Induction](https://aclanthology.org/2024.findings-acl.898) (Park & Kim, Findings 2024)
ACL