Poulomi Dey
Computer Engineering
Research Topic: Point-cloud based machine learning for classifying rare events in the AT-TPC
Research Poster Located in the Neighborhood Engagement Centers
Research Summary:
Poulomi is developing a method to efficiently identify rare fission events within a massive dataset generated by a particle accelerator experiment. To accomplish this, she combines a traditional rule-based approach with a cutting-edge machine learning technique called PointNet. By combining these methods, Poulomi has significantly improved the accuracy of detecting fission events, which is crucial for understanding nuclear reactions.
