blaise-g/msc_project

My master thesis investigates how a Natural Language Processing powered tool could be implemented to research the vast scientific literature on topics of longevity and human health, automatically extract key takeaways from relevant papers and eventually distill evidence-based personalized lifestyle recommendations given some input data.

Scientific papers automatic summarization and extraction of evidence-based lifestyle reccomendations (MSc Project)

• Author: Lorenzo Germini
• University: EPFL
• Program: MSc Life Sciences Engineering
• University Supervisor: Prof Jacques Fellay
• Industry Partner: Burgeon Labs
• Academic year: 2021/2022 (Due: August 12, 2022)

Abstract

My master project explores the implementation of an automatic abstractive text summarization pipeline with deep learning tools to extract key-takeaways from full-text research articles in the biomedical literature by distilling more concise text versions. The developed system is able to handle the complex information contained in a long scientific document given in input and subsequently produces scientific summaries according to a selected level of conciseness into a detailed or one sentence long TL;DR summary version.

A working demo of the project built with Gradio is available at https://huggingface.co/spaces/Blaise-g/summarize-biomedical-papers-long-summary-or-tldr