Determining Uncertainty in Clinical Report Writing

From REU@MU
Jump to: navigation, search

Project Background

It is important in clinical report writing to be clear in delivering diagnoses for both the patient and the doctors delivering care. When words that negate these conclusions are used in reports, it can lead to confusion and potentially lead to someone not recieving proper care. The goal of this project is to find the best way of quantifying uncertainty in clinical reports using natural language processing and machine learning to determine the quality of the report.

Goals & Milestones

  • Meet with peers and mentors
  • Review relevant resarch papers to understand problem statement
  • Look over dataset and and get a better understanding of it
  • Learn relevant technologies (e.g, Machine learning, Large Language Modeling etc)
  • Determine how to quantify and measure uncertainty
  • Apply learned technologies to dataset and start to test qualities of uncertainty
  • Determine pros/cons of different approaches to machinr learning measuring uncertainty
  • Report back to peers and mentors with paper and poster