Aiden Yun
8th Grade, The Nueva School
San Mateo, CA
A few years ago, Aiden’s mom had a breast cancer scare. Luckily, his mother did not have cancer, even though the system that diagnosed her said her chances were between 50 and 95 percent. “I was surprised by this level of confidence, or lack thereof,” he says. When he heard that lung cancer screening standards had been expanded, “I figured it was a great opportunity to do something I’d wanted to do for a long time.” Early detection is crucial for lung cancer treatment, and Aiden decided to develop an AI-powered system that would help doctors diagnose it more accurately.
Improved Accuracy in Lung Cancer Detection Integrated into a Medical Viewer
View Poster
Project Background
Aiden collected lung computed tomography (CT) scans from three different public datasets, the National Lung Screening Trial, the Lung Image Database Consortium and Lung Nodule Analysis 2016. Using these, he trained a computer model using a type of self-supervised learning called Masked Auto Encoding, which trains the computer model by hiding certain pieces of data and requires the model to fill in the blanks. Aiden then tested his model against the data in his lung CT dataset and compared his own model’s performance to current models in use.
“My method outperforms two current state-of-the-art self-supervised learning methods in lung CT scan classification, demonstrating its promising potential,” he says. Aiden also researched how radiologists do their work, and incorporated his model with a DICOM viewer, used for medical images. Eventually, he hopes to get a partnership and bring his model to the clinic.
Beyond the Project
Aiden plays competitive squash, has been a member of the USA national team, and was ranked first in the country in the boys <13 division. He also loves taking math classes at the Art of Problem Solving school and participating in math and computer science competitions. He hopes to be a computer scientist and found his own company. “I think people are too complacent towards the world’s issues,” he says. “We need to take decisive action before it gets worse.”
