arXiv subjects

Shubing Liu

Publications and source records attributed to Shubing Liu.

4 recordsLinked to original sources

arXiv2026

Detailed investigation on the Geant4 simulation of alpha particles in liquid scintillator detector

Purpose: Alpha particles from $^{214}$Po, $^{212}$Po, and $^{210}$Po are increasingly used to calibrate large liquid scintillator detectors, but Geant4 simulations of their energy deposition require detailed validation. Methods: We simulate alpha particles in a 4 m cube of LAB-based liquid scintillator with Geant4 11.4.2 and apply Birks quenching per step. We study step-size, production-threshold, Birks-constant, and electromagnetic-model dependence, and use Geant4-DNA to examine keV-scale electron track structure. Results: Delta-electron-associated fluctuations dominate the simulated response. Across the tested explicit-delta settings, the variance of the quenched-response contribution from secondary tracks is 1.01--1.12 times the total variance; a negative primary--secondary covariance reduces the total. The simulated quenching-curve shape is compatible with SNO+ data within current uncertainties, with $χ^2/\mathrm{ndf}\simeq0.14$--$0.17$ for $k_B=0.007$--$0.010$ g cm$^{-2}$ MeV$^{-1}$. Geant4-DNA gives strong electron-track curvature, with tortuosity decreasing from about 6 at 200 eV to 1.6 at 10 keV. Conclusion: Delta-electron fluctuations are the dominant contribution to the intrinsic resolution in the tested configurations. The production threshold has a larger effect than the electromagnetic model: Penelope and Livermore agree within about 3% in L/E, whereas Standard gives higher L/E, especially at low thresholds.

hep-ex↗

arXiv2026

A unified quenching model in Geant 4 simulation for $α$, proton and electron particles in liquid scintillator detectors

The liquid scintillator (LS) detectors are widely used in reactor neutrino experiments. To precisely measure the oscillated neutrino spectrum, it is crucial to understand the LS energy response, especially the quenching effect. Numerous bench-top measurements have been conducted on quenching effects for $α$ particles, protons, and electrons. These results have typically been described by Birks' law, but with different Birks' coefficients required for different particle species. In this study, we find that if more secondary electrons are allowed to be generated in Geant4 simulations, the bench-top results for all particles can be well fitted using a single Birks' coefficient of about $0.013~{\rm g\,cm^{-2}\,MeV^{-1}}$. The underlying reason is that a large fraction of the primary energy is deposited through the generation of $δ$-electrons with energies below 4~keV, which should be tracked separately due to their different quenching behavior compared to primary particles. This study provides a unified framework for comparing Birks' coefficients from different bench-top measurements and helps experiments like JUNO and SNO+ better tune their simulations for precision energy measurements.

hep-ex↗

arXiv2024

An Improved Dung Beetle Optimizer for Random Forest Optimization

To improve the convergence speed and optimization accuracy of the Dung Beetle Optimizer (DBO), this paper proposes an improved algorithm based on circle mapping and longitudinal-horizontal crossover strategy (CICRDBO). First, the Circle method is used to map the initial population to increase diversity. Second, the longitudinal-horizontal crossover strategy is applied to enhance the global search ability by ensuring the position updates of the dung beetle. Simulations were conducted on 10 benchmark test functions, and the results demonstrate that the improved algorithm performs well in both convergence speed and optimization accuracy. The improved algorithm is further applied to the hyperparameter selection of the Random Forest classification algorithm for binary classification prediction in the retail industry. Various combination comparisons prove the practicality of the improved algorithm, followed by SHapley Additive exPlanations (SHAP) analysis.

math.OC↗

arXiv2024

Enhanced Self-Checkout System for Retail Based on Improved YOLOv10

With the rapid advancement of deep learning technologies, computer vision has shown immense potential in retail automation. This paper presents a novel self-checkout system for retail based on an improved YOLOv10 network, aimed at enhancing checkout efficiency and reducing labor costs. We propose targeted optimizations to the YOLOv10 model, by incorporating the detection head structure from YOLOv8, which significantly improves product recognition accuracy. Additionally, we develop a post-processing algorithm tailored for self-checkout scenarios, to further enhance the application of system. Experimental results demonstrate that our system outperforms existing methods in both product recognition accuracy and checkout speed. This research not only provides a new technical solution for retail automation but offers valuable insights into optimizing deep learning models for real-world applications.

cs.CV↗