About Ethan Zhou
Ethan worked on the mathematical theory behind a type of machine learning called online learning and how the learner performs when predicting something very unpredictable. His work applies both when the prediction depends on a single number and when it depends on many.
Online Learning of Smooth Functions
View poster
Ethan Zhou, 18, of Vienna, worked on a type of machine learning for his Regeneron Science Talent Search mathematics project. Ethan’s work is especially useful for algorithms that predict events that are naturally revealed over time. For example, the program may be tasked with predicting Tuesday’s weather and then predicting Wednesday’s weather once it learns whether it was right about Tuesday. Over time, the predictions get better and better as the algorithm learns from its mistakes. Ethan investigated how well these algorithms can perform when confronted with something very unpredictable. He developed new bounds in both the single-variable case, when the prediction depends on a single number like temperature, and multi-variable cases, when a prediction involves more than one numerical input, like temperature, humidity and wind speed.
At McLean High School, Ethan is a three-time USA Math Olympiad qualifier and heads the math league, where he helped double the size of the club. He is also a prize-winning pianist, a skill he has honed for more than a decade. The son of Jia Hong and Guangming Zhou, Ethan has also tutored at the Hope Chinese School for five years.
Beyond the Project
FUN FACTS: Ethan was one of about 100 students selected for the Atlas Fellowship’s summer program where attendees from around the world develop big-picture thinking on science, philosophy and the future.
