Hilary Utaegbulam | Searcharxiv

arXiv2026

Facilitating neutron and energy reconstruction in neutrino events using the direction of tagged neutrons

To date, accelerator neutrino experiments have had only limited success including neutrons in the neutrino event reconstruction event by event. Produced neutrons often interact in a detector leaving an energy deposition that does not correlate strongly with the kinetic energy of the neutron. This work explores the inclusion of the direction of a neutron tagged by such an energy deposition in concert with the missing transverse momentum of the reconstructed particles to determine the approximate kinetic energy of the neutron and modify the reconstruction of the energy of the incoming neutrino. The technique significantly increases the neutron kinetic energy estimation relative to one that assigns the neutron kinetic energy by enforcing transverse momentum balance alone. When included in the neutrino energy calculation, the neutrino energy resolution is improved and the reconstructed energy is distributed more symmetrically around the true value. The technique shows promise and might be used to good effect to analyze data taken with experiments able to tag neutron energy deposits with good neutron direction resolution such as the T2K near detector and the liquid argon detectors in the SBN program and DUNE.

hep-ex↗

arXiv2026

Physics at the Edge: Benchmarking Quantisation Techniques and the Edge TPU for Neutrino Interaction Recognition

This work presents a comprehensive benchmark of different quantisation techniques for convolutional neural networks applied to neutrino interaction recognition. Utilising simulation for a generic liquid argon time-projection chamber, models are quantised and then deployed on the Google Coral Edge TPU. Models are tasked with recognising which neutrino interaction is simulated in the image between neutral current, muon-neutrino charged current, and electron-neutrino charged current. Four Keras models are tested, and accuracy is measured across two different pipelines: using post-training integer quantisation and quantisation-aware training. Inference speed is benchmarked against an AMD EPYC 7763 CPU and NVIDIA A100 GPU. A study of the energy consumption is also presented, with attention to potential costs and environmental issues. Results show that, among the four models tested, accuracy degradation is limited and, in particular, Inception V3 presents almost no accuracy degradation across the two quantisation and deployment pipelines. The speed of the edge TPU is comparable to that of the CPU, and one order of magnitude slower than the GPU. Moreover, the energy consumption of all models deployed on the edge TPU is several orders of magnitude lower than that of the CPU and GPU. In the energy consumption-latency parameter space, CPU, GPU, and edge TPU performances can be clearly separated. This paper explores possible future integrations of edge AI technologies with neutrino physics.

physics.ins-det↗