Scientific Text Mining and Information Extraction Using Deep Learning and Large Language Models

Scientific Text Mining and Information Extraction Using Deep Learning and Large Language Models

Thursday 15/05 14:00-15:30h

Workshop Space A

Session Description

Researchers begin projects by analyzing numerous scientific works to understand the state of the art, a time-consuming process that increasingly requires automation due to the growing volume of publications. In response, many have developed models to automate scientific text mining (STM) and scientific information extraction (SIE) for building retrieval-augmented generation (RAG) systems. However, these efforts often overlook the diverse needs of users, which should be considered in large-scale projects. Additionally, key information must be retrieved from entire texts, a process known as document-level processing, which requires further research due to performance limitations. To address these challenges, this workshop will introduce methods for developing STM and SIE models to extract and retrieve key information within specific domains using RAG. Participants will also learn to design personalized STM, SIE, and RAG systems to build knowledge bases and visualize as knowledge graphs.