Presented By: Megan Lim, Carle Illinois College of Medicine, University of Illinois at Urbana-Champaign
Co-Authors: Adam Cross, University of Illinois College of Medicine at Peoria
Pengcheng Jiang, University of Illinois Urbana-Champaign
Jimeng Sun, Carle Illinois College of Medicine, University of Illinois at Urbana-Champaign

Purpose 
Medical education entails the daunting task of comprehending vast amounts of information to be synthesized and understood in a limited timeframe. To alleviate this challenge, our project aims to create a web-based tool that harnesses the power of a large language model (LLM) to construct and visually present a clinical knowledge graph tailored to the specific context of the textbook being read.  

Methods 
The application allows the user to upload any learning resource such as textbooks, research articles, and lectures. After indicating pages or concepts of interest, a query is sent to ChatGPT, a LLM developed by OpenAI, and a knowledge graph that visually displays connections within the text is generated. The graph can be further modified by the user through addition or deletion of specified nodes and edges as well as saved and reloaded for future access.   

Results 
After launching the tool to first and second year medical students at the Carle Illinois College of Medicine, there are currently 63 active users of the site. Feedback indicated that the most helpful graphs were generated from research articles and textbooks, such as Robbins and Cotran Pathologic Basis of Diseases, rather than United States Medical Licensing Examination board preparation resources such as First Aid. It was also noted that the graphs offered a different perspective of the material and highlighted connections between concepts that were not immediately evident upon reading the text.  

Conclusion 
This tool can enable students to engage in active learning by practicing explaining the connections made in the graphs that offer a potentially different outlook than their own. The vision is to construct a library of knowledge graphs shared between medical schools and students across the nation through collaborative effort.