Deep-learning-based low-energy trigger algorithms for the Hyper-Kamiokande experiment
Abstract
Modern machine learning techniques have become increasingly important in particle physics because of their powerful pattern-recognition capabilities, including in real-time data acquisition where stringent runtime constraints apply. This paper details the performance of deep-learning-based trigger algorithms for a large water Cherenkov detector such as Hyper-Kamiokande, aimed at low-energy neutrino events (below 7 MeV). The performance of custom neural-network supervised classifiers is shown alongside two anomaly-detection approaches trained solely on detector noise: a pure autoencoder and a model based on Manifold Projection-Diffusion Recovery. The supervised model shows signal identification efficiencies of 76.7% for single electrons of 3 MeV kinetic energy, significantly exceeding signal efficiencies obtained from a traditional hit-count-based trigger of 26.4%, while the Manifold Projection-Diffusion Recovery approach reaches 35.4% at the same operating point. Runtime evaluations on GPU yield per-window inference latencies well below the millisecond scale.
Versions
Related
- Exploring the Supersymmetric Grand Unified Theories with Gauge Mediation at the Future Proton-Proton Colliders and Hyper-Kamiokande Experiment
- A Deep Neural Network based Level-1 Trigger for the Magnetic Monopole Search in the SCEP Experiment
- Improving the CMS High Level Trigger tracking at the HL-LHC with novel and evolved heterogeneous algorithms
- Search for the 16O(ppp) → 13C π+ π+ e+ Decay Mode in Super-Kamiokande Using Machine Learning Techniques
- Gain-Switched Sub-Nanosecond Mid-Infrared Cr:ZnSe Lasers: Experiment, Modeling, and Energy Scaling
BibTeX
@misc{lachner2026deeplearningbased,
title = {{Deep-learning-based low-energy trigger algorithms for the Hyper-Kamiokande experiment}},
author = {Katharina Lachner and Saúl Alonso-Monsalve and Benjamin Richards and Davide Sgalaberna},
howpublished = {arXiv},
year = {2026},
month = oct,
doi = {10.48550/arxiv.2605.31391},
eprint = {2605.31391},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2605.31391},
note = {CC BY 4.0},
}
RIS
TY - UNPB
TI - Deep-learning-based low-energy trigger algorithms for the Hyper-Kamiokande experiment
AU - Katharina Lachner
AU - Saúl Alonso-Monsalve
AU - Benjamin Richards
AU - Davide Sgalaberna
PY - 2026
DA - 2026/10/16
PB - arXiv
DO - 10.48550/arxiv.2605.31391
UR - https://arxiv.org/abs/2605.31391
AB - Modern machine learning techniques have become increasingly important in particle physics because of their powerful pattern-recognition capabilities, including in real-time data acquisition where stringent runtime constraints apply. This paper details the performance of deep-learning-based trigger algorithms for a large water Cherenkov detector such as Hyper-Kamiokande, aimed at low-energy neutrino events (below 7 MeV). The performance of custom neural-network supervised classifiers is shown alongside two anomaly-detection approaches trained solely on detector noise: a pure autoencoder and a model based on Manifold Projection-Diffusion Recovery. The supervised model shows signal identification efficiencies of 76.7% for single electrons of 3 MeV kinetic energy, significantly exceeding signal efficiencies obtained from a traditional hit-count-based trigger of 26.4%, while the Manifold Projection-Diffusion Recovery approach reaches 35.4% at the same operating point. Runtime evaluations on GPU yield per-window inference latencies well below the millisecond scale.
ER -