DeepJet: A Machine Learning Environment for High-energy Physics

Swapneel Mehta, Mauro Verzetti, Jan Kieseler, Markus Stoye

Keywords: deep learning, high-energy physics, tensorflow

TL;DR: A deep learning framework that simplifies definition, training, and evaluation of Tensorflow models in a production environment for Physicists at CERN.

Abstract: The DeepJet Framework is used for applying cutting-edge practices in deep learning to supervised learning problems in high-energy physics. Originally envisaged to support jet-flavour tagging and classification, it has grown to encompass a range of use-cases as it underwent a transformation into a multi-purpose tool for physics analysis at the Compact Muon Solenoid (CMS) Experiment. This paper illustrates the motivation behind the development of DeepJet, its features, architecture, and workflow.

Decision: reject