In high school, he was part of a research project hosted by the University of Nebraska called the Cosmic Ray Observatory Project (CROP). The project's goal, which was the topic of his senior project, was building cosmic ray detectors and putting them on the roofs of Nebraska high schools. In college, he worked with Professor Jim Alexander in the Cornell University CMS group. They developed and deployed a new method for measuring the top quark mass with early data from the Large Hadron Collider.
In graduate school, his Ph.D. co-adviser Ariel Schwartzman had been interested in using jets to look for new particles. Ben became fascinated by the internal structure of jets and the wealth of information their substructure can tell us about the quantum properties of the strong force. His interest has coincided with a renaissance in studying the strong force. Over the last approximately ten years, experimentalists and theorists have made amazing progress in studying the radiation pattern inside jets.
Ben said many of the calibration and machine learning tools he developed likely have broader applicability across science and industry. In particular, his research has led to new methods for combining AI with computer simulations and training machine learning methods on data without labels, that is pre-classified into categories (imagine training a neural network to distinguish cat and dog pictures where the training dataset doesn’t have the animals labeled).
One key question he hopes to solve over the next 10 to 20 years is how we can ensure that we fully exploit the data from the Large Hadron Collider (LHC)? Researchers continue to pore over the data from the LHC, but all of the explorations so far only take advantage of low-dimensional information and are guided by theory. Ben believes that a complementary approach that is more exploratory, within theoretically reasonable bounds, could also yield important insight. Ben wonders if the new physics can only be observed through subtle correlations in high dimensions. The growing field of AI-enabled anomaly detection for fundamental physics will likely play a central role in this research.