Text Summarization using Pegasus Model

Benny, Anu (2024) Text Summarization using Pegasus Model. Masters thesis, Dublin, National College of Ireland.

Preview	PDF (Master of Science) Download (531kB) | Preview
Preview	PDF (Configuration Manual) Download (578kB) | Preview

Abstract

As digital content grows at an unprecedented rate, dealing with information overload has become a vital concern. In order to improve summarization from conversational text, this project uses the SAMSum dataset to fine-tune the Pegasus model. Our goal was to strengthen the model’s capacity to produce clear and succinct summaries, which led to a significant improvement in the Rouge score—a crucial indicator of summarization quality. Because of the vast amount of text data that is generated every day—including emails and chat logs—effective summarization is essential. The Rouge-1 score was successfully raised by our fine-tuning, indicating better summarization accuracy. To further highlight the model’s adaptability, I increased its capacity to summarize YouTube videos and extract and summarize text from photos. These enhancements provide useful advantages for more effective summarization and are in line with recent developments in NLP. In order to get over current constraints and improve the model’s wider applicability, future work will concentrate on better optimizing it for intricate, multimodal inputs.

Item Type:	Thesis (Masters)
Supervisors:	Name Email Jain, Mayank UNSPECIFIED
Uncontrolled Keywords:	Summarization; finetune; Pegasus
Subjects:	Q Science > QA Mathematics > Electronic computers. Computer science T Technology > T Technology (General) > Information Technology > Electronic computers. Computer science H Social Sciences > HM Sociology > Information Science > Communication P Language and Literature > P Philology. Linguistics > Computational linguistics. Natural language processing
Divisions:	School of Computing > Master of Science in Artificial Intelligence
Depositing User:	Ciara O'Brien
Date Deposited:	18 Jun 2025 10:40
Last Modified:	18 Jun 2025 10:47
URI:	https://norma.ncirl.ie/id/eprint/7902

Actions (login required)

View Item