Deep learning framework for enhanced neutrino reconstruction of single-line events in the ANTARES telescope

Albert A, Alves S, André M, Ardid M, Ardid S, Aubert JJ, Aublin J, Baret B, Basa S, Becherini Y, Belhorma B, Benfenati F, Bertin V, Biagi S, Boumaaza J, Bouta M, Bouwhuis MC, Brânzaş H, Bruijn R, Brunner J, Busto J, Caiffi B, Calvo D, Campion S, Capone A, Carenini F, Carr J, Carretero V, Cartraud T, Celli S, Cerisy L, Chabab M, Cherkaoui El Moursli R, Chiarusi T, Circella M, Coelho JA, Coleiro A, Coniglione R, Coyle P, Creusot A, Díaz AF, De Martino B, Distefano C, Di Palma I, Donzaud C, Dornic D, Drouhin D, Eberl T, Eddymaoui A, van Eeden T, van Eijk D, El Hedri S, El Khayati N, Enzenhöfer A, Fermani P, Ferrara G, Filippini F, Fusco L, Gagliardini S, García-Méndez J, Gatius Oliver C, Gay P, Geißelbrecht N, Glotin H, Gozzini R, Gracia Ruiz R, Graf K, Guidi C, Haegel L, van Haren H, Heijboer AJ, Hello Y, Hennig L, Hernández-Rey JJ, Hößl J, Huang F, Illuminati G, Jisse-Jung B, de Jong M, de Jong P, Kadler M, Kalekin O, Katz U, Kouchner A, Kreykenbohm I, Kulikovskiy V, Lahmann R, Lamoureux M, Lazo A, Lefèvre D, Leonora E, Levi G, Le Stum S, Loucatos S, Manczak J, Marcelin M, Margiotta A, Marinelli A, Martínez-Mora JA, Migliozzi P, Moussa A, Muller R, Navas S, Nezri E, Ó Fearraigh B, Oukacha E, Păun AM, Păvălaş GE, Peña-Martínez S, Perrin-Terrin M, Piattelli P, Poirè C, Popa V, Pradier T, Randazzo N, Real D, Riccobene G, Romanov A, Sánchez Losa A, Saina A, Salesa Greus F, Samtleben DF, Sanguineti M, Sapienza P, Schüssler F, Seneca J, Spurio M, Stolarczyk T, Taiuti M, Tayalati Y, Vallage B, Vannoye G, Van Elewyck V, Viola S, Vivolo D, Wilms J, Zavatarelli S, Zegarelli A, Zornoza JD, Zúñiga J (2026)

Publication Type: Journal article

Publication year: 2026

Journal

Book Volume: 7

Journal Issue: 3

DOI: 10.1088/2632-2153/ae5d84

Abstract

We present the N-fit algorithm designed to improve the reconstruction of neutrino events detected by a single line of the ANTARES underwater telescope, usually associated with low energy neutrino events (~100 GeV). N-Fit is a neural network model that relies on deep learning and combines several advanced techniques in machine learning—deep convolutional layers, mixture density output layers, and transfer learning (TL). This framework divides the reconstruction process into two dedicated branches for each neutrino event topology—tracks and showers—composed of sub-models for spatial estimation—direction and position—and energy inference, which later on are combined for event classification. Regarding the direction of single-line (SL) events, the N-Fit algorithm significantly refines the estimation of the zenithal angle, and delivers reliable azimuthal angle predictions that were previously unattainable with traditional χ2-fit methods. Improving on energy estimation of SL events is a tall order; N-Fit benefits from TL to efficiently integrate key characteristics, such as the estimation of the closest distance from the event to the detector. NFit also takes advantage from TL in event topology classification by freezing convolutional layers of the pretrained branches. Tests on Monte Carlo simulations and data demonstrate a significant reduction in mean and median absolute errors across all reconstructed parameters. The improvements achieved by N-Fit highlight its potential for advancing multimessenger astrophysics and enhancing our ability to probe fundamental physics beyond the Standard Model using SL events from ANTARES data.

Authors with CRIS profile

Involved external institutions

Université de Strasbourg (UDS) France (FR) Institute for Corpuscular Physics / Instituto de Física Corpuscular (IFIC) Spain (ES) Polytechnic University of Catalonia / Universitat Politècnica de Catalunya · BarcelonaTech (UPC) / Universidad Politécnica de Cataluña Spain (ES) Polytechnic University of Valencia / Universidad Politécnica de Valencia Spain (ES) Aix-Marseille University / Aix-Marseille Université France (FR) Laboratoire AstroParticule et Cosmologie APC / Astroparticle and Cosmology Laboratory France (FR) The National Center for Nuclear Energy, Sciences and Techniques (CNESTEN) Morocco (MA) National Institute for Nuclear Physics / Istituto Nazionale di Fisica Nucleare (INFN) Italy (IT) Mohammed V University (UMVA) / جامعة محمد الخامس‎ Morocco (MA) Mohammed First University / Université Mohammed Premier / جامعة محمد الأول Morocco (MA) FOM Foundation Netherlands (NL) National Institute for Laser, Plasma & Radiation Physics (INFLPR) Romania (RO) Royal Netherlands Institute for Sea Research (NIOZ) Netherlands (NL) UMR Géoazur - Campus Azur du CNRS France (FR) Julius-Maximilians-Universität Würzburg Germany (DE) Mediterranean Institute of Oceanography (MIO) / Institut méditerranéen d'océanologie France (FR) Università degli Studi di Salerno Italy (IT) University of Toulon / Université de Toulon (UTLN) France (FR) Cadi Ayyad University / جامعة القاضي عياض‎ Morocco (MA) Universidad de Granada Spain (ES) Institute of Research into the Fundamental Laws of the Universe / Institut de recherche sur les lois fondamentales de l'univers (IRFU) France (FR)

How to cite

APA:

Albert, A., Alves, S., André, M., Ardid, M., Ardid, S., Aubert, J.J.,... Zúñiga, J. (2026). Deep learning framework for enhanced neutrino reconstruction of single-line events in the ANTARES telescope. Machine Learning: Science and Technology, 7(3). https://doi.org/10.1088/2632-2153/ae5d84

MLA:

Albert, A., et al. "Deep learning framework for enhanced neutrino reconstruction of single-line events in the ANTARES telescope." Machine Learning: Science and Technology 7.3 (2026).

BibTeX: Download