Colloquium – Mariel Pettee – Machine Learning for High-Energy Particle Physics

Speaker: Dr. Mariel Pettee (LBNL)

Title: Machine Learning for High-Energy Particle Physics

Abstract:

10 years ago, in the summer of 2012, CERN announced the discovery of the Higgs boson, marking the end of a half-century quest to complete the Standard Model. Later that year, a deep neural network called AlexNet outperformed other models by an astounding margin on a popular image recognition competition, arguably marking the start of the deep learning revolution of the last decade. In the 10 years since that pivotal year, the fields of high-energy particle physics and deep learning have been inextricably entwined. Machine learning is now an essential ingredient in the experimental process at CERN, enabling us to take data at unprecedented rates and helping us probe these mountains of data for hints of new physics. In this talk, I will present a framework for understanding how machine learning has been successfully deployed for high-energy particle physics as well as what physicists can uniquely contribute to the field of machine learning in return. I will also argue that future innovation in this space will necessarily come from interdisciplinary work — by drawing connections between our specific challenges and those from other datasets and domains.