中法核工程与技术学院核声论坛（总第229期）

About the speaker：

Qiyu Yan is a Ph.D. in experimental high-energy physics at the University of Chinese Academy of Sciences (UCAS), advised by Prof. Yangheng Zheng. He is also a visiting student at the University of Warwick, working with Prof. Xianguo Lu. His research focuses on neutrino-nucleus interactions, cross-section modeling, and systematic uncertainty evaluations for next-generation neutrino experiments.

Within the JUNO collaboration, Qiyu serves as a task lead for the GENIE and NuWro event generator efforts in the atmospheric neutrino group. His key contributions include implementing the Ghent Hybrid pion production model into the NuWro generator and developing systematic uncertainty propagation frameworks based on the Professor 2 method. Additionally, he actively explores modern computational techniques in physics, such as Markov Chain Monte Carlo (MCMC) methods and GPU-accelerated fitting and inference. 

Abstract:

Neutrino physics has entered the precision era, where theoretical uncertainties in neutrino-nucleus interactions have become a dominant systematic limitation for flagship measurements—the neutrino mass ordering, the CP-violating phase, and proton decay searches. This talk presents a coherent study of nuclear effects spanning bottom-up model development to top-level experimental sensitivity projections. I will first discuss the implementation of the Ghent Hybrid single-pion production model into the NuWro event generator, which is rigorously validated against transverse kinematic imbalance data from MINERvA and MicroBooNE. Following this, I will present a systematic assessment of nuclear medium effects using the GiBUU transport framework, revealing intriguing tensions among current experimental datasets in their preference for different medium-effect strengths. Building on these foundational models, I will then demonstrate how nuclear effects propagate into proton decay searches in next-generation water Cherenkov detectors, where initial-state Fermi motion, final-state interactions, and atmospheric-neutrino backgrounds are self-consistently quantified within the same GiBUU framework. Finally, turning to the JUNO liquid-scintillator detector, I will present an atmospheric neutrino oscillation analysis that clarifies the origin of mass-ordering sensitivity and the critical role of flavor identification. To address the computational challenges in these large-scale fits, I will introduce a novel systematic-error propagation framework based on SVD-assisted parameter-space interpolation, designed to efficiently handle the high-dimensional nuclear-uncertainty space and enable robust sensitivity projections for JUNO's atmospheric neutrino program.