Bereiche | Tage | Auswahl | Suche | Aktualisierungen | Downloads | Hilfe

T: Fachverband Teilchenphysik

T 77: Deep Learning III

T 77.5: Vortrag

Donnerstag, 28. März 2019, 17:00–17:15, H06

A deep learning based search for a heavy CP-even Higgs boson in dileptonic H → WW decays with the CMS experiment — •Peter Fackeldey¹ and Dennis Roy² — ¹III. Physikalisches Institut A, RWTH Aachen University — ²III. Physikalisches Institut B, RWTH Aachen University

A promising model beyond the Standard Model is the Minimal Supersymmetric Standard Model (MSSM), which is commonly parameterized in the Higgs sector by tanβ and m_A. As in every 2HDM, five different Higgs bosons are predicted. Especially the decay of the heavy scalar Higgs boson into two W bosons is very sensitive to low values of tanβ and m_A. Standard Model background processes are a challenge in this region. These backgrounds are modelled using data driven methods, whose performances heavily rely on the purity of their associated control regions. In the last few years Deep Learning showed remarkable progress and success in high energy physics. We present a multi-class classification strategy with deep neural networks, which increases the purity in the signal regions and in control regions for SM background processes. This strategy minimizes systematic uncertainties and thus improves the limits in an unexplored region of the MSSM parameter space in the search for H → WW.