High School Students Advised by Computer Science Faculty Win Big at National Competitions

Assistant Professor Abhinav Shrivastava and Assistant Professor John Dickerson’s mentoring leads high school students to the top 300 scholars in this year’s Regeneron Science Talent Search, the nation’s oldest and most prestigious competition.
Descriptive image for High School Students Advised by Computer Science Faculty Win Big at National Competitions

Michelle Tang, advised by Assistant Professor Abhinav Shrivastava, and Zach Zhao, advised by Assistant Professor John Dickerson, were selected as semifinalists in the  2020 Regeneron Science Talent Search.  The Regeneron STS is the oldest and arguably most prestigious high school competition in the United States.  On top of this, Zach Zhao’s project went on to win the ACM/CSTA Cutler-Bell Prize, established in 2015 by DEC/Microsoft alumni David Cutler and Gordon Bell, which recognizes students who pursue computing challenges beyond the traditional classroom environment.

Both Tang and Zhao are high school seniors in the Science, Mathematics, and Computer Science Magnet Program at Montgomery Blair High School, located in nearby Montgomery County, Maryland.

Michelle Tang has been working on independent research projects outside of school since seventh grade. She has been participating in the regional Montgomery County Science Fair every year since she was in middle school. “Participating in Regeneron STS and having my work be recognized on a national level has been a dream that I’ve had ever since I began my research journey in middle school,” she said.

Tang’s research project focuses on how medical images can be used and analyzed to help doctors diagnose autism patients. Brain disorders like autism are often difficult to accurately identify. Tools like functional magnetic resonance imaging (fMRI) allow physicians to perform visual analysis on brain scans to help diagnose neurological diseases, but as the analysis is manual, it  is susceptible to human error, said Tang.
Her work aims to build systems by using computer vision, that can learn to identify anomalies in medical images, and then predict the correct diagnosis based on them.

As a kid, Zach Zhao had a passion for STEM and did several projects related to machine learning and computer science within and outside of school. He did his first comprehensive research project under the guidance of Prof. Dickerson last summer. “I want to share my research with the broader community and being selected as a Semi-Finalist confirms my contribution and serves as a milestone for my research,” said Zhao. 

Zhao’s project focused on finding compatible pairs of patients and kidney donors from a pool of participants by exploring a novel data-driven approach to find matchings by using a combination of deep neural networks (specifically graph neural networks, or GNNs) and tree search to learn from precomputed matchings.

The trained GNN understands the structure of a pool of patients and donors and make an educated guess of what a good matching could be and the tree search algorithm refines this educated guess by searching through additional possibilities thereby enabling quality matching,  Zhao explained.

Both Tang and Zhao had high praise for their computer science mentors.

“Working with Dr. Shrivastava has been a huge honor and an amazing learning opportunity” says Tang.  “He taught me everything from hyperparameter optimization to how to self-correct my experimental design in the research process. He always made time to answer my questions and guide me throughout the course of my project,” she said.

“I really enjoyed working with Professor Dickerson, who fully supported me during the research,” said Zhao.  “Despite his busy schedule, Professor Dickerson was always there to help me and it is because of his support that I successfully finished my first research paper, which was accepted to the AAAI Student Abstract Program and also got me Regeneron Semi-Finalist,” said Zhao.

Both Tang and Zhao plan to further their education in computer science and continue their passion for doing real-world applications of their original research.

The Regeneron Science Talent Search provides students a national stage to present original research. The Regeneron Science Talent Search scholars were chosen based on their exceptional research skills, commitment to academics and innovative thinking.  The chosen scholars and their schools will be awarded $2,000 each.

The ACM/CSTA Cutler-Bell Prize promotes the field of computer science and empowers students to pursue computing challenges beyond the traditional classroom environment. In 2015, David Cutler and Gordon Bell established the award. Each Cutler-Bell Prize winner receives a $10,000 cash prize.

The Department welcomes comments, suggestions and corrections.  Send email to editor [-at-] cs [dot] umd [dot] edu.