17-18 September 2018
Alan Turing Institute, London
Europe/London timezone

Direct optimisation of the discovery significance when training neural networks to search for new physics in particle colliders

18 Sep 2018, 14:30
15m

Speaker

Dr Adam Elwood

Description

Dr. Adam Elwood

Abstract:"We introduce two new loss functions designed to directly optimise the statistical significance of the expected number of signal events when training neural networks to classify events as signal or background in the scenario of a search for new physics at a particle collider. The loss functions are designed to directly maximise commonly used estimates of the statistical significance, s/√(s+b), and the Asimov estimate, Z_A. We consider their use in a toy SUSY search with 30~fb^(−1) of 14~TeV data collected at the LHC. In the case that the search for the SUSY model is dominated by systematic uncertainties, it is found that the loss function based on Z_A can outperform the binary cross entropy in defining an optimal search region."
arXiv: https://arxiv.org/abs/1806.00322

Bio: I’m a postdoctoral researcher at the DESY particle physics laboratory in Hamburg and a member of the CMS collaboration. My interests are in the applications of machine learning techniques to searches for Beyond the Standard Model (BSM) physics, particularly supersymmetry and dark matter. I completed my PhD at Imperial College London with a search for supersymmetry in the all hadronic final state at the CMS experiment.

Presentation Materials

There are no materials yet.