11/12/19: Deep Learning at the LHC with Professor Javier Duarte

Professor Javier Duarte visited us this week from University of California, San Diego. He discussed his research with machine learning in high energy physics. Professor Duarte uses deep neural networks and graph neural networks to separate background from signal in Higgs to bb decays. He also discussed using generative neural networks in particle flow, an algorithm that clusters hits in different parts of the detector for one particle. Another application of machine learning in physics that Professor Duarte mentioned was using autoencoders to search for anomalies in our data that could point to new physics. Finally, Professor Duarte concluded by discussing how deep learning is implemented in the trigger level at the LHC in the form of FPGAs.

Professor Durate’s slides can be found here