Teen Innovator Wins $40,000 for Machine-Learning Breakthrough on Particle Collisions

Seth Nabat, an 18-year-old senior from Winnetka, California, secured tenth place and a $40,000 award in the 2026 Regeneron Science Talent Search, the nation’s longest-running high-school science competition. His entry features a three-part machine-learning architecture: one network encodes symmetry to speed up collision approximations, a second unconstrained network captures camera and measurement glitches, and a third extracts patterns from those errors.

Tests showed the combined model could navigate noisy data while retaining computational efficiency, offering physicists a new tool to study symmetry-breaking in particle physics. Nabat also serves as co-captain of his school’s varsity debate team and mentors elementary and middle-school students through the UCLA Math Circle, reflecting his commitment to education. His parents, Robyn Brook-Nabat and Scott Nabat, have supported his pursuits. The prize underscores the role of youth-led innovation in advancing scientific research and may influence future studies in physics and cosmology.