No Cover Image

Conference Paper/Proceeding/Abstract 543 views 96 downloads

GlossGPT: GPT for Word Sense Disambiguation using Few-shot Chain-of-Thought Prompting

Deshan Sumanathilaka Orcid Logo, Nicholas Micallef Orcid Logo, Julian Hough Orcid Logo

Procedia Computer Science, Volume: 257, Pages: 785 - 792

Swansea University Authors: Deshan Sumanathilaka Orcid Logo, Nicholas Micallef Orcid Logo, Julian Hough Orcid Logo

  • 68937.VOR.pdf

    PDF | Version of Record

    © 2025 The Authors. This is an open access article under the CC BY-NC-ND license.

    Download (329.64KB)

Check full text

DOI (Published version): 10.1016/j.procs.2025.03.101

Abstract

Lexical ambiguity is a major challenge in computational linguistic tasks, as limitations in proper sense identification lead to inefficient translation and question answering. General-purpose Large Language Models (LLMs) are commonly utilized for Natural Language Processing (NLP) tasks. However, uti...

Full description

Published in: Procedia Computer Science
ISSN: 1877-0509
Published: Elsevier BV 2025
Online Access: Check full text

URI: https://cronfa.swan.ac.uk/Record/cronfa68937
first_indexed 2025-02-21T11:52:11Z
last_indexed 2025-04-29T04:40:11Z
id cronfa68937
recordtype SURis
fullrecord <?xml version="1.0"?><rfc1807><datestamp>2025-04-28T10:27:30.6505466</datestamp><bib-version>v2</bib-version><id>68937</id><entry>2025-02-21</entry><title>GlossGPT: GPT for Word Sense Disambiguation using Few-shot Chain-of-Thought Prompting</title><swanseaauthors><author><sid>2fe44f0c1e7d845dc21bb6b00d5b2085</sid><ORCID>0009-0005-8933-6559</ORCID><firstname>Deshan</firstname><surname>Sumanathilaka</surname><name>Deshan Sumanathilaka</name><active>true</active><ethesisStudent>false</ethesisStudent></author><author><sid>1cc4c84582d665b7ee08fb16f5454671</sid><ORCID>0000-0002-2683-8042</ORCID><firstname>Nicholas</firstname><surname>Micallef</surname><name>Nicholas Micallef</name><active>true</active><ethesisStudent>false</ethesisStudent></author><author><sid>082d773ae261d2bbf49434dd2608ab40</sid><ORCID>0000-0002-4345-6759</ORCID><firstname>Julian</firstname><surname>Hough</surname><name>Julian Hough</name><active>true</active><ethesisStudent>false</ethesisStudent></author></swanseaauthors><date>2025-02-21</date><deptcode>MACS</deptcode><abstract>Lexical ambiguity is a major challenge in computational linguistic tasks, as limitations in proper sense identification lead to inefficient translation and question answering. General-purpose Large Language Models (LLMs) are commonly utilized for Natural Language Processing (NLP) tasks. However, utilizing general-purpose LLMs for specific tasks has been challenging, and fine-tuning has become a critical requirement for task specification. In this work, we craft advanced prompts with different contextual parameters to guide the model's inference towards accurate sense prediction to handle Word Sense Disambiguation (WSD). We present a few-shot Chain of Thought (COT) prompt-based technique using GPT-4-Turbo with knowledgebase as a retriever that does not require fine-tuning the model for WSD tasks and sense definitions are supported by synonyms to broaden the lexical meaning. Our approach achieves comparable performance on the SemEval and Senseval datasets. More importantly, we set a new state-of-the-art performance with the few-shot FEWS dataset, breaking through the 90% F1 score barrier.</abstract><type>Conference Paper/Proceeding/Abstract</type><journal>Procedia Computer Science</journal><volume>257</volume><journalNumber/><paginationStart>785</paginationStart><paginationEnd>792</paginationEnd><publisher>Elsevier BV</publisher><placeOfPublication/><isbnPrint/><isbnElectronic/><issnPrint/><issnElectronic>1877-0509</issnElectronic><keywords>Word Sense Disambiguation, Knowledge Base Retrieval, Large Language Models, GPT-4-Turbo, Chain of Thought Prompting</keywords><publishedDay>25</publishedDay><publishedMonth>4</publishedMonth><publishedYear>2025</publishedYear><publishedDate>2025-04-25</publishedDate><doi>10.1016/j.procs.2025.03.101</doi><url/><notes/><college>COLLEGE NANME</college><department>Mathematics and Computer Science School</department><CollegeCode>COLLEGE CODE</CollegeCode><DepartmentCode>MACS</DepartmentCode><institution>Swansea University</institution><apcterm>Other</apcterm><funders/><projectreference/><lastEdited>2025-04-28T10:27:30.6505466</lastEdited><Created>2025-02-21T11:46:03.2490240</Created><path><level id="1">Faculty of Science and Engineering</level><level id="2">School of Mathematics and Computer Science - Computer Science</level></path><authors><author><firstname>Deshan</firstname><surname>Sumanathilaka</surname><orcid>0009-0005-8933-6559</orcid><order>1</order></author><author><firstname>Nicholas</firstname><surname>Micallef</surname><orcid>0000-0002-2683-8042</orcid><order>2</order></author><author><firstname>Julian</firstname><surname>Hough</surname><orcid>0000-0002-4345-6759</orcid><order>3</order></author></authors><documents><document><filename>68937__34119__fb55d411dc2542108a5951e3f76c5469.pdf</filename><originalFilename>68937.VOR.pdf</originalFilename><uploaded>2025-04-28T10:20:11.3864882</uploaded><type>Output</type><contentLength>337556</contentLength><contentType>application/pdf</contentType><version>Version of Record</version><cronfaStatus>true</cronfaStatus><documentNotes>&#xA9; 2025 The Authors. This is an open access article under the CC BY-NC-ND license.</documentNotes><copyrightCorrect>true</copyrightCorrect><language>eng</language><licence>https://creativecommons.org/licenses/by-nc-nd/4.0</licence></document></documents><OutputDurs/></rfc1807>
spelling 2025-04-28T10:27:30.6505466 v2 68937 2025-02-21 GlossGPT: GPT for Word Sense Disambiguation using Few-shot Chain-of-Thought Prompting 2fe44f0c1e7d845dc21bb6b00d5b2085 0009-0005-8933-6559 Deshan Sumanathilaka Deshan Sumanathilaka true false 1cc4c84582d665b7ee08fb16f5454671 0000-0002-2683-8042 Nicholas Micallef Nicholas Micallef true false 082d773ae261d2bbf49434dd2608ab40 0000-0002-4345-6759 Julian Hough Julian Hough true false 2025-02-21 MACS Lexical ambiguity is a major challenge in computational linguistic tasks, as limitations in proper sense identification lead to inefficient translation and question answering. General-purpose Large Language Models (LLMs) are commonly utilized for Natural Language Processing (NLP) tasks. However, utilizing general-purpose LLMs for specific tasks has been challenging, and fine-tuning has become a critical requirement for task specification. In this work, we craft advanced prompts with different contextual parameters to guide the model's inference towards accurate sense prediction to handle Word Sense Disambiguation (WSD). We present a few-shot Chain of Thought (COT) prompt-based technique using GPT-4-Turbo with knowledgebase as a retriever that does not require fine-tuning the model for WSD tasks and sense definitions are supported by synonyms to broaden the lexical meaning. Our approach achieves comparable performance on the SemEval and Senseval datasets. More importantly, we set a new state-of-the-art performance with the few-shot FEWS dataset, breaking through the 90% F1 score barrier. Conference Paper/Proceeding/Abstract Procedia Computer Science 257 785 792 Elsevier BV 1877-0509 Word Sense Disambiguation, Knowledge Base Retrieval, Large Language Models, GPT-4-Turbo, Chain of Thought Prompting 25 4 2025 2025-04-25 10.1016/j.procs.2025.03.101 COLLEGE NANME Mathematics and Computer Science School COLLEGE CODE MACS Swansea University Other 2025-04-28T10:27:30.6505466 2025-02-21T11:46:03.2490240 Faculty of Science and Engineering School of Mathematics and Computer Science - Computer Science Deshan Sumanathilaka 0009-0005-8933-6559 1 Nicholas Micallef 0000-0002-2683-8042 2 Julian Hough 0000-0002-4345-6759 3 68937__34119__fb55d411dc2542108a5951e3f76c5469.pdf 68937.VOR.pdf 2025-04-28T10:20:11.3864882 Output 337556 application/pdf Version of Record true © 2025 The Authors. This is an open access article under the CC BY-NC-ND license. true eng https://creativecommons.org/licenses/by-nc-nd/4.0
title GlossGPT: GPT for Word Sense Disambiguation using Few-shot Chain-of-Thought Prompting
spellingShingle GlossGPT: GPT for Word Sense Disambiguation using Few-shot Chain-of-Thought Prompting
Deshan Sumanathilaka
Nicholas Micallef
Julian Hough
title_short GlossGPT: GPT for Word Sense Disambiguation using Few-shot Chain-of-Thought Prompting
title_full GlossGPT: GPT for Word Sense Disambiguation using Few-shot Chain-of-Thought Prompting
title_fullStr GlossGPT: GPT for Word Sense Disambiguation using Few-shot Chain-of-Thought Prompting
title_full_unstemmed GlossGPT: GPT for Word Sense Disambiguation using Few-shot Chain-of-Thought Prompting
title_sort GlossGPT: GPT for Word Sense Disambiguation using Few-shot Chain-of-Thought Prompting
author_id_str_mv 2fe44f0c1e7d845dc21bb6b00d5b2085
1cc4c84582d665b7ee08fb16f5454671
082d773ae261d2bbf49434dd2608ab40
author_id_fullname_str_mv 2fe44f0c1e7d845dc21bb6b00d5b2085_***_Deshan Sumanathilaka
1cc4c84582d665b7ee08fb16f5454671_***_Nicholas Micallef
082d773ae261d2bbf49434dd2608ab40_***_Julian Hough
author Deshan Sumanathilaka
Nicholas Micallef
Julian Hough
author2 Deshan Sumanathilaka
Nicholas Micallef
Julian Hough
format Conference Paper/Proceeding/Abstract
container_title Procedia Computer Science
container_volume 257
container_start_page 785
publishDate 2025
institution Swansea University
issn 1877-0509
doi_str_mv 10.1016/j.procs.2025.03.101
publisher Elsevier BV
college_str Faculty of Science and Engineering
hierarchytype
hierarchy_top_id facultyofscienceandengineering
hierarchy_top_title Faculty of Science and Engineering
hierarchy_parent_id facultyofscienceandengineering
hierarchy_parent_title Faculty of Science and Engineering
department_str School of Mathematics and Computer Science - Computer Science{{{_:::_}}}Faculty of Science and Engineering{{{_:::_}}}School of Mathematics and Computer Science - Computer Science
document_store_str 1
active_str 0
description Lexical ambiguity is a major challenge in computational linguistic tasks, as limitations in proper sense identification lead to inefficient translation and question answering. General-purpose Large Language Models (LLMs) are commonly utilized for Natural Language Processing (NLP) tasks. However, utilizing general-purpose LLMs for specific tasks has been challenging, and fine-tuning has become a critical requirement for task specification. In this work, we craft advanced prompts with different contextual parameters to guide the model's inference towards accurate sense prediction to handle Word Sense Disambiguation (WSD). We present a few-shot Chain of Thought (COT) prompt-based technique using GPT-4-Turbo with knowledgebase as a retriever that does not require fine-tuning the model for WSD tasks and sense definitions are supported by synonyms to broaden the lexical meaning. Our approach achieves comparable performance on the SemEval and Senseval datasets. More importantly, we set a new state-of-the-art performance with the few-shot FEWS dataset, breaking through the 90% F1 score barrier.
published_date 2025-04-25T05:26:50Z
_version_ 1851097781055258624
score 11.089386

Similar Items