Background rejection in NEXT using deep neural networks
Abstract
We investigate the potential of using deep learning techniques to reject background events in searches for neutrinoless double beta decay with high pressure xenon time projection chambers capable of detailed track reconstruction. The differences in the topological signatures of background and signal events can be learned by deep neural networks via training over many thousands of events. These networks can then be used to classify further events as signal or background, providing an additional background rejection factor at an acceptable loss of efficiency. The networks trained in this study performed better than previous methods developed based on the use of the same topological signatures by a factor of 1.2 to 1.6, and there is potential for further improvement.
Keywords
Cite
@article{arxiv.1609.06202,
title = {Background rejection in NEXT using deep neural networks},
author = {NEXT Collaboration and J. Renner and A. Farbin and J. Muñoz Vidal and J. M. Benlloch-Rodríguez and A. Botas and P. Ferrario and J. J. Gómez-Cadenas and V. Álvarez and C. D. R. Azevedo and F. I. G. Borges and S. Cárcel and J. V. Carrión and S. Cebrián and A. Cervera and C. A. N. Conde and J. Díaz and M. Diesburg and R. Esteve and L. M. P. Fernandes and A. L. Ferreira and E. D. C. Freitas and A. Goldschmidt and D. González-Díaz and R. M. Gutiérrez and J. Hauptman and C. A. O. Henriques and J. A. Hernando Morata and V. Herrero and B. Jones and L. Labarga and A. Laing and P. Lebrun and I. Liubarsky and N. López-March and D. Lorca and M. Losada and J. Martín-Albo and G. Martínez-Lema and A. Martínez and F. Monrabal and C. M. B. Monteiro and F. J. Mora and L. M. Moutinho and M. Nebot-Guinot and P. Novella and D. Nygren and A. Para and J. Pérez and M. Querol and L. Ripoll and J. Rodríguez and F. P. Santos and J. M. F. dos Santos and L. Serra and D. Shuman and A. Simón and C. Sofka and M. Sorel and J. F. Toledo and J. Torrent and Z. Tsamalaidze and J. F. C. A. Veloso and J. White and R. Webb and N. Yahlali and H. Yepes-Ramírez},
journal= {arXiv preprint arXiv:1609.06202},
year = {2017}
}
Comments
21 pages, 9 figures; formatting changes