Center for Theoretical Physics of the Universe (CTPU-PTC)

Spectral Analysis for Jet Substructure with Neural Network: Boosted Higgs Case

Name: Spectral Analysis for Jet Substructure with Neural Network: Boosted Higgs Case
Start: 2018-07-13T15:00:00+09:00
End: 2018-07-13T16:00:00+09:00
Location: Theory Building 4th Floor

by Dr Sung Hak Lim (KEK)

Friday 13 Jul 2018, 15:00 → 16:00 Asia/Seoul

CTPU seminar room (Theory Building 4th Floor)

CTPU seminar room

Theory Building 4th Floor

IBS Main Campus, Daejeon

Description

Non-QCD jets typically have a particular angular scale which can be used to distinguish it from QCD jets. We introduce a new strategy for jet substructure analysis using a spectral function of the angular scale and an artificial neural network. Our spectrum allows us to scan over the angular scale based on energy deposits in a highly visualizable way. The neural network then tries to find out a peculiar shape of the spectrum over particular angular scales. This approach has similar performance to the existing algorithm to find the jet substructure. We show the performance of this strategy regarding boosted Higgs boson disambiguation from QCD jets.