Reducing model bias in a deep learning classifier using domain adversarial neural networks in the MINERvA experiment

Nuruzzaman, G. N. Perdue, A. Ghosh, M. Wospakrik, F. Akbar, D. A. Andrade, M. Ascencio, L. Bellantoni, A. Bercellie, M. Betancourt, G. F.R.Caceres Vera, T. Cai, M. F. Carneiro, J. Chaves, D. Coplowe, H. Da Motta, G. A. Díaz, J. Felix, L. Fields, R. FineA. M. Gago, R. Galindo, T. Golan, R. Gran, J. Y. Han, D. A. Harris, D. Jena, J. Kleykamp, M. Kordosky, X. G. Lu, E. Maher, W. A. Mann, C. M. Marshall, K. S. McFarland, A. M. McGowan, B. Messerly, J. Miller, J. K. Nelson, C. Nguyen, A. Norrick, Nuruzzaman Nuruzzaman, A. Olivier, R. Patton, M. A. Ramírez, R. D. Ransome, H. Ray, L. Ren, D. Rimal, D. Ruterbories, H. Schellman, C. J.Solano Salinas, H. Su, S. Upadhyay, E. Valencia, J. Wolcott, B. Yaeggy, S. Young

Research output: Contribution to journalArticlepeer-review

7 Scopus citations

Abstract

We present a simulation-based study using deep convolutional neural networks (DCNNs) to identify neutrino interaction vertices in the MINERvA passive targets region, and illustrate the application of domain adversarial neural networks (DANNs) in this context. DANNs are designed to be trained in one domain (simulated data) but tested in a second domain (physics data) and utilize unlabeled data from the second domain so that during training only features which are unable to discriminate between the domains are promoted. MINERvA is a neutrino-nucleus scattering experiment using the NuMI beamline at Fermilab. A-dependent cross sections are an important part of the physics program, and these measurements require vertex finding in complicated events. To illustrate the impact of the DANN we used a modified set of simulation in place of physics data during the training of the DANN and then used the label of the modified simulation during the evaluation of the DANN. We find that deep learning based methods offer significant advantages over our prior track-based reconstruction for the task of vertex finding, and that DANNs are able to improve the performance of deep networks by leveraging available unlabeled data and by mitigating network performance degradation rooted in biases in the physics models used for training.

Original languageEnglish (US)
Article numberP11020
JournalJournal of Instrumentation
Volume13
Issue number11
DOIs
StatePublished - Nov 26 2018

Bibliographical note

Funding Information:
This document was prepared by the MINERνA collaboration using the resources of the Fermi National Accelerator Laboratory (Fermilab), a U.S. Department of Energy, Office of Science, HEP User Facility. Fermilab is managed by Fermi Research Alliance, LLC (FRA), acting under Contract No. DE-AC02-07CH11359, which included the MINERνA construction project. The research here was aslo sponsored by the Laboratory Directed Research and Development Program of Oak Ridge National Laboratory, managed by UT-Battelle, LLC, for the U.S. Department of Energy. This research used resources of the Oak Ridge Leadership Computing Facility at the Oak Ridge National Laboratory, which is supported by the Office of Science of the U.S. Department of Energy under Contract No. DE-AC05-00OR22725. Construction support also was granted by the United States National Science Foundation under Award PHY-0619727 and by the University of Rochester. Additional support for participating scientists was provided by NSF and DOE (U.S.A.) by CAPES and CNPq (Brazil), by CoNaCyT (Mexico), by Proyecto Basal FB 0821, CONICYT PIA ACT1413, Fondecyt 3170845 and 11130133 (Chile), by PIIC (DGIP-UTFSM), by CONCYTEC, DGI-PUCP and IDI/IGI-UNI (Peru), by Latin American Center for Physics (CLAF), by RAS and the Russian Ministry of Education and Science (Russia), and by the National Science Centre of Poland, grant number DEC-2017/01/X/ST2/00128. We thank the MINOS Collaboration for use of its near detector data. Finally, we thank the staff of Fermilab for support of the beamline and the detector.

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

Keywords

  • Analysis and statistical methods
  • Neutrino detectors
  • Pattern recognition, cluster finding, calibration andfitting methods

Fingerprint

Dive into the research topics of 'Reducing model bias in a deep learning classifier using domain adversarial neural networks in the MINERvA experiment'. Together they form a unique fingerprint.

Cite this