arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2605.13627physics.ins-detphysics.data-an
收藏

SINAPSE: 一种用于精确且可解释的中子-γ甄别的轻量级深度学习框架

SINAPSE: A lightweight deep learning framework for accurate and explainable neutron-$γ$ discrimination

University of Caen Normandy(诺曼底卡昂大学);ENSICAEN(法国国立高等工程学院);CNRS/IN2P3, LPC Caen UMR6534(法国国家科学研究中心/法国国家核子与粒子物理研究所,卡昂低能物理联合研究实验室);CEA, DAM, DIF(法国原子能和替代能源委员会,达姆研究所,惯性聚变部门);Université Paris-Saclay(巴黎萨克雷大学);CEA, Laboratoire Matière en Conditions Extrêmes(法国原子能和替代能源委员会,极端条件物质实验室)
Thomas Carreau, Adrien Matta, Owen Syrett, Benoît Mauss, David Etasse, Cyril Lenain, Pierre Morfouace, Julien Taieb, David Regnier, Patrick Copp, Matthew Devlin… 展开作者

Thomas Carreau, Adrien Matta, Owen Syrett, Benoît Mauss, David Etasse, Cyril Lenain, Pierre Morfouace, Julien Taieb, David Regnier, Patrick Copp, Matthew Devlin, Charlène Surault, Jason Surbrook

AI总结:

提出SINAPSE轻量级深度学习框架,采用双分支架构(一维卷积自编码器降噪+分类器),通过随机增强模拟低电荷条件,实现低信噪比下精确且可解释的中子-γ甄别,性能优于传统方法,且决策基于物理有意义的脉冲形状特征。

摘要:

传统上,有机闪烁体中的中子-γ甄别依赖于飞行时间(ToF)选择和脉冲形状甄别(PSD)等技术。然而,在低电荷区域,由于信噪比(SNR)较差,通过图形切割进行粒子识别仍然具有挑战性。在这项工作中,我们提出了SINAPSE,一个轻量级深度学习框架,用于低电荷区域中精确且可解释的中子-γ甄别。该框架采用双分支架构,将用于波形降噪的一维卷积自编码器与用于粒子识别的分类器相结合。对高SNR波形应用随机增强以模拟低电荷条件,从而能够稳健地外推到传统PSD标签不可靠的区域。我们表明,与传统的数字信号处理技术相比,SINAPSE实现了优越的降噪性能,并输出校准良好的概率,与传统图形切割一致。最后,我们应用SHAP(SHapley Additive exPlanations)值表明模型决策由物理有意义的脉冲形状特征驱动,确认了与已建立的PSD原理的一致性。

英文摘要:

Traditionally, neutron-$γ$ discrimination in organic scintillators relies on techniques such as time-of-flight (ToF) selection and pulse-shape discrimination (PSD). However, particle identification through graphical cuts remains challenging in the low-charge regime due to poor signal-to-noise ratios (SNR). In this work, we propose SINAPSE, a lightweight deep learning framework for accurate and explainable neutron-$γ$ discrimination in the low-charge regime. The framework employs a dual-branch architecture that combines a 1-dimensional convolutional autoencoder for waveform denoising with a classifier for particle identification. Random augmentations are applied to high-SNR waveforms to simulate low-charge conditions, enabling robust extrapolation into regimes where conventional PSD labels are unreliable. We show that SINAPSE achieves superior denoising performance compared to conventional digital signal processing techniques, and outputs well-calibrated probabilities, consistent with traditional graphical cuts. Finally, we apply SHAP (SHapley Additive exPlanations) values to show that model decisions are driven by physically meaningful pulse-shape features, confirming consistency with established PSD principles.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑