实验粒子物理计算研讨会（2023）

Session

Machine Learning

Jul 17, 2023, 2:50 PM

Conveners

Machine Learning

• There are no conveners in this block

Machine Learning

• Manqi RUAN (Institute of High Energy Physics, Beijing, China)

Machine Learning

• There are no conveners in this block

9. Reconstruction of Neutrino Event in TRIDENT Based on TridentNet

Cen Mo

7/17/23, 2:50 PM

机器学习

The tRoplcal DEep-sea Neutrino Telescope (TRIDENT) is a next-generation neutrino detector located in the South China Sea. High computational efficiency is required for event reconstruction methods in order to calculate the incident particle's direction and energy. In a typical neutrino event, less than 1% of photosensors are hit, making Graph Neural Networks particularly well-suited for their...

17. Machine Learning Based Tracking Reconstruction in the Muon g-2 Experiment at Fermilab

Bingzhi Li

7/17/23, 3:15 PM

机器学习

报告

The first result of the Fermilab muon g-2 experiment shows that the anomalous magnetic moment of the muon has 4.2 standard deviation between the experimental and theoretical result, which provides a strong evidence for the new physics beyond the Standard Model. The tracking reconstruction plays an important role in many aspects of the experiment. In the Run1 analyses, both the speed and...

32. ParticleNet and its application at CEPC Jet Flavor Tagging

Mr 永峰 朱 (北京大学)

7/17/23, 4:10 PM

机器学习

ParticleNet, a customized neural network architecture based on Dynamic Graph Convolutional Neural Network, has achieved significant improvements over all existing methods in two jet tagging tasks in proton-proton collisions: top tagging and quark-gluon tagging. Currently, it is widely used in physics analyses conducted by CMS. In this study, we employ ParticleNet in the CEPC and analyze its...

29. Machine Learning in Track reconstruction at DarkSHINE Experiment

Yulei Zhang

7/17/23, 4:35 PM

机器学习

The DarkSHINE Experiment is a proposed fixed-target operation using an electron beam to search for dark photons by measuring the missing momentum. The measurement relies on the efficient use of tagging and recoil trackers to measure the electron energy pre- and post-target. We present a comparative analysis of conventional methodologies and contemporary machine-learning techniques for tracking...

20. PandaX-III实验中粒子径迹特征分析方法与灵敏度估算

Tao Li (Sun Yat-sen University)

7/18/23, 3:15 PM

机器学习

对马约拉纳中微子的研究是当前粒子物理领域探索超出标准模型新物理的热点研究方向，无中微子双贝塔衰变（NLDBD）是实验上可以确认中微子马约拉纳属性的稀有核衰变。PandaX-III合作组致力于打造具有国际竞争力的百公斤靶质量实验，采用基于气体微结构探测器技术的高压气氙时间投影室来寻找Xe-136的NLDBD过程，其最显著优势在于能够通过带电粒子径迹特征进行信号本底鉴别，进而大幅提高实验对NLDBD的探测灵敏度。本报告将从PandaX-III实验中带电粒子径迹特征分析入手，介绍粒子径迹重建、信号本底鉴别、事例顶点重建等方法，推动PandaX-III朝着零本底实验条件发展。