Speaker
Benjamin Nachman
(LBNL)
Description
Deep neural networks (DNNs) have revolutionized many areas of science and technology. In this talk, we will discuss cutting edge developments in DNNs for high energy physics, using jet physics (including calorimeter showers) as an example that has attracted significant recent attention. Domain specific challenges require new techniques to make full use of the algorithms. A key focus is on understanding how and what the algorithms learn. DNN techniques are demonstrated for classification, regression, and generation. In addition to providing powerful baseline performance, we show how to train complex models directly on data and to generate sparse stacked images with non-uniform granularity.
Primary authors
Benjamin Nachman
(LBNL)
Eric Metodiev
(MIT)
Francesco Rubbo
(SLAC National Accelerator Laboratory)
Lucio Dery
(Stanford)
Luke de Oliveira
(LBNL)
Matt Schwartz
(Harvard)
Michela Paganini
(Yale)
komiske Patrick
(MIT)