An introduction to Neural Networks in HEP and developments for their application to new physics searches

by Mr Adam Elwood (DESY)

Monday 28 May 2018, 14:00 → 15:00 Europe/Berlin

Seminar Room 03 (DESY Hamburg)

In this seminar I’ll give a brief introduction to neural networks, introducing the concept and field specific jargon. I will then give an overview of some of the different neural network architectures and some ways they are being put to use in HEP. I will then go on to outline a new method to directly optimise for the discovery significance when training neural networks to search for new physics in particle colliders. Here I introduce a new loss function, based on the Asimov estimate of significance, which performs better than the binary cross entropy in the case of signal models that are dominated by systematic uncertainties.

Slides 180522_studentSeminarNNs.pdf