PreprintCC BY 4.0arXiv
Machine-learning-based neutrino-event generation constrained by experimental measurements
Abstract
We present a generative surrogate for charged-current νμ-nucleus scattering that can be calibrated to measured differential cross sections without event-level experimental samples. Starting from a theoretical interaction model, we fine-tune the surrogate to a MicroBooNE measurement and show that this also improves the description of observables not used in the fit. The calibrated model can also be applied to measurements from different neutrino fluxes without retuning, while retaining event-level predictions of outgoing-muon kinematics. Our results show that experimental cross sections can directly constrain generative event models while preserving their applicability across different observables and beam configurations.
Related
- CP-violation or Nuclear Excitation: Reviewing the Role of Neutrino Interaction Model Uncertainties on Accelerator-Based Neutrino Oscillation Measurements
- Goodness-of-fit for multi-distribution neutrino cross-section measurements with shared events
- An effective topology-transfer nuisance for Genie-based neutrino-argon semi-inclusive analyses: a methodological proposal with MicroBooNE-constrained diagnostic
- Right-handed neutrino dark matter consistent with the generation of baryon number asymmetry
- Measurements of transverse kinematic imbalance in muon neutrino charged-current interactions on argon with a single proton and no pions in the final state with SBND
BibTeX
@misc{bonilla2026machinelearningbased,
title = {{Machine-learning-based neutrino-event generation constrained by experimental measurements}},
author = {Jose Luis Bonilla and Krzysztof M. Graczyk and Rwik Dharmapal Banerjee and Beata E. Kowal and Hemant Prasad and Jan T. Sobczyk},
howpublished = {arXiv},
year = {2026},
month = oct,
doi = {10.48550/arxiv.2610.20870},
eprint = {2610.20870},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2610.20870},
note = {CC BY 4.0},
}
RIS
TY - UNPB
TI - Machine-learning-based neutrino-event generation constrained by experimental measurements
AU - Jose Luis Bonilla
AU - Krzysztof M. Graczyk
AU - Rwik Dharmapal Banerjee
AU - Beata E. Kowal
AU - Hemant Prasad
AU - Jan T. Sobczyk
PY - 2026
DA - 2026/10/14
PB - arXiv
DO - 10.48550/arxiv.2610.20870
UR - https://arxiv.org/abs/2610.20870
AB - We present a generative surrogate for charged-current ν_μ-nucleus scattering that can be calibrated to measured differential cross sections without event-level experimental samples. Starting from a theoretical interaction model, we fine-tune the surrogate to a MicroBooNE measurement and show that this also improves the description of observables not used in the fit. The calibrated model can also be applied to measurements from different neutrino fluxes without retuning, while retaining event-level predictions of outgoing-muon kinematics. Our results show that experimental cross sections can directly constrain generative event models while preserving their applicability across different observables and beam configurations.
ER -