Deep neural networks for energy and position reconstruction in EXO-200

S. Delaquis, M. J. Jewell, I. Ostrovskiy, M. Weber, T. Ziegler, J. Dalmasson, L. J. Kaufman, T. Richards, J. B. Albert, G. Anton, I. Badhrees, P. S. Barbeau, R. Bayerlein, D. Beck, V. Belov, M. Breidenbach, T. Brunner, G. F. Cao, W. R. Cen, C. ChambersB. Cleveland, M. Coon, A. Craycraft, W. Cree, T. Daniels, M. Danilov, S. J. Daugherty, J. Daughhetee, J. Davis, A. Der Mesrobian-Kabakian, R. Devoe, J. Dilling, A. Dolgolenko, M. J. Dolinski, W. Fairbank, J. Farine, S. Feyzbakhsh, P. Fierlinger, D. Fudenberg, R. Gornea, G. Gratta, C. Hall, E. V. Hansen, D. Harris, J. Hoessl, P. Hufschmidt, M. Hughes, A. Iverson, A. Jamil, A. Johnson, A. Karelin, T. Koffas, S. Kravitz, R. Krücken, A. Kuchenkov, K. S. Kumar, Y. Lan, D. S. Leonard, G. S. Li, S. Li, C. Licciardi, Y. H. Lin, R. Maclellan, T. Michel, B. Mong, D. Moore, K. Murray, O. Njoya, A. Odian, A. Piepke, A. Pocar, F. Retière, A. L. Robinson, P. C. Rowson, S. Schmidt, A. Schubert, D. Sinclair, A. K. Soma, V. Stekhanov, M. Tarka, J. Todd, T. Tolba, V. Veeraraghavan, J. L. Vuilleumier, M. Wagenpfeil, A. Waite, J. Watkins, L. J. Wen, U. Wichoski, G. Wrede, Q. Xia, L. Yang, Y. R. Yen, O. Ya Zeldovich

Research output: Contribution to journalArticlepeer-review

35 Scopus citations

Abstract

We apply deep neural networks (DNN) to data from the EXO-200 experiment. In the studied cases, the DNN is able to reconstruct the relevant parameters - total energy and position - directly from raw digitized waveforms, with minimal exceptions. For the first time, the developed algorithms are evaluated on real detector calibration data. The accuracy of reconstruction either reaches or exceeds what was achieved by the conventional approaches developed by EXO-200 over the course of the experiment. Most existing DNN approaches to event reconstruction and classification in particle physics are trained on Monte Carlo simulated events. Such algorithms are inherently limited by the accuracy of the simulation. We describe a unique approach that, in an experiment such as EXO-200, allows to successfully perform certain reconstruction and analysis tasks by training the network on waveforms from experimental data, either reducing or eliminating the reliance on the Monte Carlo.

Original languageEnglish
Article numberP08023
JournalJournal of Instrumentation
Volume13
Issue number8
DOIs
StatePublished - Aug 29 2018

Bibliographical note

Publisher Copyright:
© 2018 IOP Publishing Ltd and Sissa Medialab.

Funding

We thank Adam Coates (Baidu) for helpful and encouraging discussions and Evan Racah (NERSC, LBNL) for support of deep learning applications at NERSC at the beginning of this work. We thank the Erlangen Regional Computing Center (RRZE) for the compute resources and support. We thank the Deutsche Forschungsgemeinschaft (DFG) for the support of this study. EXO-200 data analysis and simulation uses resources of the National Energy Research Scientific Computing Center (NERSC), which is supported by the Office of Science of the U.S. Department of Energy under Contract No. DE-AC02-05CH11231. We gratefully acknowledge the support of Nvidia Corporation with the donation of the Titan Xp GPU used for this research. We thank Adam Coates (Baidu) for helpful and encouraging discussions and Evan Racah (NERSC, LBNL) for support of deep learning applications at NERSC at the beginning of this work. We thank the Erlangen Regional Computing Center (RRZE) for the compute resources and support. We thank the Deutsche Forschungsgemeinschaft (DFG) for the support of this study. EXO-200 data analysis and simulation uses resources of the National Energy Research Scientific Computing Center (NERSC), which is supported by the Office of Science of the U.S. Department of Energy under Contract No. DE-AC02-05CH11231. EXO-200 is supported by DOE and NSF in the U.S., NSERC in Canada, SNF in Switzerland, IBS in Korea, RFBR in Russia, DFG in Germany, and CAS and ISTCP in China. We gratefully acknowledge the KARMEN Collaboration for supplying the cosmic-ray veto detectors, and the WIPP for their hospitality.

FundersFunder number
U.S. Department of Energy Chinese Academy of Sciences Guangzhou Municipal Science and Technology Project Oak Ridge National Laboratory Extreme Science and Engineering Discovery Environment National Science Foundation National Energy Research Scientific Computing Center National Natural Science Foundation of China1654495
U.S. Department of Energy Chinese Academy of Sciences Guangzhou Municipal Science and Technology Project Oak Ridge National Laboratory Extreme Science and Engineering Discovery Environment National Science Foundation National Energy Research Scientific Computing Center National Natural Science Foundation of China
U.S. Department of Energy Oak Ridge National Laboratory U.S. Department of Energy National Science Foundation National Energy Research Scientific Computing CenterDE-AC02-05CH11231
U.S. Department of Energy Oak Ridge National Laboratory U.S. Department of Energy National Science Foundation National Energy Research Scientific Computing Center
National Science Foundation Office of International Science and Engineering
Lawrence Berkeley National Laboratory
Natural Sciences and Engineering Research Council of Canada
Deutsche ForschungsgemeinschaftEXO-200
Deutsche Forschungsgemeinschaft
Schweizerischer Nationalfonds zur Förderung der Wissenschaftlichen Forschung
Russian Foundation for Basic Research
Chinese Academy of Sciences
International Science and Technology Cooperation Programme

    Keywords

    • Analysis and statistical methods
    • Double-beta decay detectors
    • Pattern recognition, cluster finding, calibration and fitting methods
    • Time projection chambers

    ASJC Scopus subject areas

    • Instrumentation
    • Mathematical Physics

    Fingerprint

    Dive into the research topics of 'Deep neural networks for energy and position reconstruction in EXO-200'. Together they form a unique fingerprint.

    Cite this