Skip To Main Content

Custom Class: header-container

Custom Class: header-utility-container

Custom Class: header-breadcrumb

 

Seniors Help Predict Pneumonia in Patients During Encounter Experience

During their Senior Encounter with Bill King, Sr. Product Manager, Digital Health & Precision Medicine at Johns Hopkins University, Alex Boulay '23 and Kaikai Du '23 used ChatGPT to write a Python script to predict whether a patient has pneumonia.

Using an AI-inspired software development technique called 'prompt engineering', they asked ChatGPT to generate code to train an algorithm (ML) to process chest x-rays using an open source dataset. The result is a system that can predict pneumonia with >85% accuracy. The project is publicly available on Github so others interested in medical imaging informatics have a place to start working on the future of medicine and software development.

“Heading into Senior Encounter, I was a bit uncertain since I didn't have much prior experience in data science, especially within the medical field. However, I was pleasantly surprised by how many topics piqued my interest,” said Boulay. “This Encounter was an incredibly valuable experience for me. It not only enhanced my problem-solving skills but also taught me how to network effectively and promote myself. If I could give any advice to underclassmen or rising seniors, it would be to fully embrace the Encounter and give it your all. By maximizing your efforts and engagement, you can truly reap the benefits that Encounter has to offer.”

“I was really surprised at the direction this Encounter took since I expected it would be more of a shadowing experience instead of the hands-on project we undertook these past two weeks. I'm glad we got to do this project since I've gained valuable experience in working in a more professional setting, and it allowed for us to learn about something we might have only started doing in college,” said Du. “Our mentor, Mr. Bill King, was also incredibly supportive of us, which we are very grateful for.”

Precision Medicine suggested that the team submit their work for consideration to the Conference on Machine Intelligence in Medical Imaging, which will take place in October at Hopkins.

Congratulations, Alex and Kaikai!


 

More News and Views from Roland Avenue and Beyond