Speaker
Description
Neutrino interaction cross sections below 100 GeV are important for current and next-generation neutrino oscillation experiments. Future accelerator-based experiments such as SHiP and the planned Forward Physics Facility will be able to probe neutrino interactions in this energy regime. Reliable theoretical predictions of neutrino cross sections are essential. In this region, deep inelastic scattering (DIS) cross sections are influenced by nucleon structure functions at low momentum transfer ($Q^2$). For $Q^2$ < 1 GeV$^2$, in particular, structure functions derived from perturbative QCD based parton distribution functions are not applicable, which requires alternative approaches.
We present data-driven low $Q^2$ structure functions obtained using machine learning method with theoretical constraints. We employ a neural network and train on charged lepton and neutrino scattering data. To describe physical behavior in the low x and low $Q^2$ limits, we incorporate theoretical guidance from Regge theory and from the PCAC (partially conserved axial-vector current). We also present neutrino cross sections evaluated using the resulting structure functions.