Below are internal research opportunities advertised specifically for Geisel medical students:
Posted 12/10
Center for Technology and Behavior Health
Contact Paul Barr, PhD (paul.j.barr@dartmouth.edu)
We are looking for medical students to help fine-tune a Large Language Model (LLM) for clinical annotations.
Project: This project aims to train an LLM to annotate clinical visit transcripts to identify text segments related to Medications, Medical Conditions, Symptoms, Tests, Treatments, and Procedures. This model will be integrated into an existing web application (HealthPAL), identifying these segments to help patients navigate recordings of their clinic visits.
Task: Audit the model’s initial outputs on at least 5 transcripts to help improve accuracy and relevance. The codebook is still being finalized, so some flexibility will be needed as definitions and categories evolve.
There may be a chance for involvement in future publications related to this work.