3D track finding for MicroBooNE’s deep learning based event reconstruction chain
Adrien Hourlier
MIT
Lundi 12/11/2018, 11h00-12h00
Bat 141, salle André Berthelot, CEA Paris-Saclay

MicroBooNE is a Liquid Argon Time Projection Chamber (LArTPC) neutrino experiment on the Booster Neutrino Beamline at the Fermi National Accelerator Laboratory, with an 85-tonne active mass.

One of MicroBooNE's primary physics goals is to investigate the excess of electron neutrino events seen by MiniBooNE in the [200-600] MeV range.

MicroBooNE will constrain the intrinsic electron neutrino component of the beam by measuring the muon neutrino spectrum.

Our low-energy excess analysis makes use of deep learning algorithms applied to the high-resolution images provided by the MicroBooNE LArTPC.

I will present a novel 3D event reconstruction based on computer vision tools and a stochastic search algorithm, yielding a 2.5% energy resolution for 1mu1p muon neutrino interactions in the [200-1500] MeV range.

I will then present validation studies verifying the good agreement of our simulation to our muon neutrino data.

Contact : Fabrice BALLI

 

Retour en haut