Author:
Li S.,Ostrovskiy I.,Li Z.,Yang L.,Al Kharusi S.,Anton G.,Barbeau P.S.,Badhrees I.,Beck D.,Belov V.,Bhatta T.,Breidenbach M.,Brunner T.,Cao G.F.,Cen W.R.,Chambers C.,Cleveland B.,Coon M.,Craycraft A.,Daniels T.,Darroch L.,Daugherty S.J.,Davis J.,Delaquis S.,Der Mesrobian-Kabakian A.,DeVoe R.,Dilling J.,Dolgolenko A.,Dolinski M.J.,Echevers J.,Fairbank W.,Fairbank D.,Farine J.,Feyzbakhsh S.,Fierlinger P.,Fu Y.S.,Fudenberg D.,Gautam P.,Gornea R.,Gratta G.,Hall C.,Hansen E.V.,Hoessl J.,Hufschmidt P.,Hughes M.,Iverson A.,Jamil A.,Jessiman C.,Jewell M.J.,Johnson A.,Karelin A.,Kaufman L.J.,Koffas T.,Krücken R.,Kuchenkov A.,Kumar K.S.,Lan Y.,Larson A.,Lenardo B.G.,Leonard D.S.,Li G.S.,Licciardi C.,Lin Y.H.,MacLellan R.,McElroy T.,Michel T.,Mong B.,Moore D.C.,Murray K.,Njoya O.,Nusair O.,Odian A.,Perna A.,Piepke A.,Pocar A.,Retière F.,Robinson A.L.,Rowson P.C.,Runge J.,Schmidt S.,Sinclair D.,Skarpaas K.,Soma A.K.,Stekhanov V.,Tarka M.,Thibado S.,Todd J.,Tolba T.,Totev T.I.,Tsang R.,Veenstra B.,Veeraraghavan V.,Vogel P.,Vuilleumier J.-L.,Wagenpfeil M.,Watkins J.,Weber M.,Wen L.J.,Wichoski U.,Wrede G.,Wu S.X.,Xia Q.,Yahne D.R.,Yen Y.-R.,Zeldovich O.Ya.,Ziegler T.
Abstract
Abstract
Generative Adversarial Networks trained on samples of
simulated or actual events have been proposed as a way of generating
large simulated datasets at a reduced computational cost. In this
work, a novel approach to perform the simulation of photodetector
signals from the time projection chamber of the EXO-200 experiment
is demonstrated. The method is based on a Wasserstein Generative
Adversarial Network — a deep learning technique allowing for
implicit non-parametric estimation of the population distribution
for a given set of objects. Our network is trained on real
calibration data using raw scintillation waveforms as input. We find
that it is able to produce high-quality simulated waveforms an order
of magnitude faster than the traditional simulation approach and,
importantly, generalize from the training sample and discern salient
high-level features of the data. In particular, the network
correctly deduces position dependency of scintillation light
response in the detector and correctly recognizes dead photodetector
channels. The network output is then integrated into the EXO-200
analysis framework to show that the standard EXO-200 reconstruction
routine processes the simulated waveforms to produce energy
distributions comparable to that of real waveforms. Finally, the
remaining discrepancies and potential ways to improve the approach
further are highlighted.
Subject
Mathematical Physics,Instrumentation