Information Extraction from Contracts Using Large Language Models

Geerligs, Cornelis (2024) Information Extraction from Contracts Using Large Language Models. Master's Thesis / Essay, Artificial Intelligence.

Preview	Text mAI2024CornelisG.pdf Download (2MB) | Preview
	Text Toestemming Geerligs.pdf Restricted to Registered users only Download (208kB)

Abstract

This thesis explores the usability of Large Language Models (LLMs) to automate intricate contract reporting tasks, focusing on filling in the Digital Operational Resilience Act (DORA) compliance template and extracting obligations from contracts. Structured experiments evaluate the capability of LLMs to identify and extract relevant information from legal texts and format this information to meet specific compliance and reporting standards. The first experiment achieved an average accuracy of 97.71% in filling out the DORA compliance template. The second experi- ment, involving the extraction of contractual obligations, showed a variable performance with the highest accuracy at 70.56%. The findings demonstrate the potential of advanced language mod- els to reshape legal document analysis and management, reducing human error and labor costs in legal processes. The research also identifies the strengths and limitations of LLMs in extract- ing precise and contextually relevant information, suggesting areas for further development and optimization to enhance their applicability in the legal domain.

Item Type:	Thesis (Master's Thesis / Essay)
Supervisor name:	Dhali, M.A.
Degree programme:	Artificial Intelligence
Thesis type:	Master's Thesis / Essay
Language:	English
Date Deposited:	06 Sep 2024 09:00
Last Modified:	06 Sep 2024 09:00
URI:	https://fse.studenttheses.ub.rug.nl/id/eprint/34210

Actions (login required)

View Item