Leveraging GPT Models For Semantic Table Annotation

verfasst von: Jean Petit Bikim, Carick Atezong, Azanzi Jiomekong, Allard Oelen, Gollam Rabby, Jennifer D’Souza, Sören Auer
Abstract: This paper outlines our contribution to the Accuracy Track and the Semantic Table Interpretation (STI) & Large Language Models (LLMs) track of the Semantic Web Challenge on Tabular Data to Knowledge Graph Matching (SemTab). Our approach involves using LLMs to address the various tasks presented in the challenge. Specifically, we employed zero-shot and few-shot prompting techniques for most of the tasks, which facilitated the LLMs ability to interpret and annotate tabular data with minimal prior training. For the Column Property Annotation (CPA) task, we took a different approach by applying a set of predefined rules, tailored to the structure of each dataset. Our method achieved notable results, with an f1-score exceeding 0.92, demonstrating the effectiveness of LLMs in tackling the SemTab challenge. These results suggest that LLMs hold significant capabilities as a robust solution for semantic table annotation and knowledge graph matching, highlighting their potential to advance the field of semantic web technologies.
Organisationseinheit(en): Forschungszentrum L3S
Externe Organisation(en): Technische Informationsbibliothek (TIB) Leibniz-Informationszentrum Technik und Naturwissenschaften und Universitätsbibliothek
University of Yaounde I
Typ: Aufsatz in Konferenzband
Seiten: 43-53
Anzahl der Seiten: 11
Publikationsdatum: 03.01.2025
Publikationsstatus: Veröffentlicht
Peer-reviewed: Ja
ASJC Scopus Sachgebiete: Allgemeine Computerwissenschaft
Elektronische Version(en): https://ceur-ws.org/Vol-3889/paper3.pdf (Zugang: Offen)

BibTeX

@inproceedings{e4c30472363b4b3c99659490cdfaad4e,
title = "Leveraging GPT Models For Semantic Table Annotation",
abstract = "This paper outlines our contribution to the Accuracy Track and the Semantic Table Interpretation (STI) & Large Language Models (LLMs) track of the Semantic Web Challenge on Tabular Data to Knowledge Graph Matching (SemTab). Our approach involves using LLMs to address the various tasks presented in the challenge. Specifically, we employed zero-shot and few-shot prompting techniques for most of the tasks, which facilitated the LLMs ability to interpret and annotate tabular data with minimal prior training. For the Column Property Annotation (CPA) task, we took a different approach by applying a set of predefined rules, tailored to the structure of each dataset. Our method achieved notable results, with an f1-score exceeding 0.92, demonstrating the effectiveness of LLMs in tackling the SemTab challenge. These results suggest that LLMs hold significant capabilities as a robust solution for semantic table annotation and knowledge graph matching, highlighting their potential to advance the field of semantic web technologies.",
keywords = "Knowledge Graph, Large Language Models, Prompt Engineering, Semantic Table Annotation, Semantic Table Interpretation, SemTab, Tabular Data",
author = "Bikim, {Jean Petit} and Carick Atezong and Azanzi Jiomekong and Allard Oelen and Gollam Rabby and Jennifer D{\textquoteright}Souza and S{\"o}ren Auer",
note = "Publisher Copyright: {\textcopyright} 2024 Copyright for this paper by its authors.; 2024 Semantic Web Challenge on Tabular Data to Knowledge Graph Matching, SemTab 2024 ; Conference date: 13-11-2024 Through 13-11-2024",
year = "2025",
month = jan,
day = "3",
language = "English",
series = "CEUR workshop proceedings",
publisher = "CEUR Workshop Proceedings",
pages = "43--53",
booktitle = "Semantic Web Challenge on Tabular Data to Knowledge Graph Matching 2024",
}