Zhipeng Yao, Xingtao Huang, Teng Li, Weidong Li, Tao Lin, Jiaheng Zou
{"title":"基于变分量子分类器的 BESIII μ介子/离子识别","authors":"Zhipeng Yao, Xingtao Huang, Teng Li, Weidong Li, Tao Lin, Jiaheng Zou","doi":"arxiv-2408.13812","DOIUrl":null,"url":null,"abstract":"In collider physics experiments, particle identification (PID), i. e. the\nidentification of the charged particle species in the detector is usually one\nof the most crucial tools in data analysis. In the past decade, machine\nlearning techniques have gradually become one of the mainstream methods in PID,\nusually providing superior discrimination power compared to classical\nalgorithms. In recent years, quantum machine learning (QML) has bridged the\ntraditional machine learning and the quantum computing techniques, providing\nfurther improvement potential for traditional machine learning models. In this\nwork, targeting at the $\\mu^{\\pm} /\\pi^{\\pm}$ discrimination problem at the\nBESIII experiment, we developed a variational quantum classifier (VQC) with\nnine qubits. Using the IBM quantum simulator, we studied various encoding\ncircuits and variational ansatzes to explore their performance. Classical\noptimizers are able to minimize the loss function in quantum-classical hybrid\nmodels effectively. A comparison of VQC with the traditional multiple layer\nperception neural network reveals they perform similarly on the same datasets.\nThis illustrates the feasibility to apply quantum machine learning to data\nanalysis in collider physics experiments in the future.","PeriodicalId":501181,"journal":{"name":"arXiv - PHYS - High Energy Physics - Experiment","volume":"40 1","pages":""},"PeriodicalIF":0.0000,"publicationDate":"2024-08-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":"{\"title\":\"Muon/Pion Identification at BESIII based on Variational Quantum Classifier\",\"authors\":\"Zhipeng Yao, Xingtao Huang, Teng Li, Weidong Li, Tao Lin, Jiaheng Zou\",\"doi\":\"arxiv-2408.13812\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"In collider physics experiments, particle identification (PID), i. e. the\\nidentification of the charged particle species in the detector is usually one\\nof the most crucial tools in data analysis. In the past decade, machine\\nlearning techniques have gradually become one of the mainstream methods in PID,\\nusually providing superior discrimination power compared to classical\\nalgorithms. In recent years, quantum machine learning (QML) has bridged the\\ntraditional machine learning and the quantum computing techniques, providing\\nfurther improvement potential for traditional machine learning models. In this\\nwork, targeting at the $\\\\mu^{\\\\pm} /\\\\pi^{\\\\pm}$ discrimination problem at the\\nBESIII experiment, we developed a variational quantum classifier (VQC) with\\nnine qubits. Using the IBM quantum simulator, we studied various encoding\\ncircuits and variational ansatzes to explore their performance. Classical\\noptimizers are able to minimize the loss function in quantum-classical hybrid\\nmodels effectively. A comparison of VQC with the traditional multiple layer\\nperception neural network reveals they perform similarly on the same datasets.\\nThis illustrates the feasibility to apply quantum machine learning to data\\nanalysis in collider physics experiments in the future.\",\"PeriodicalId\":501181,\"journal\":{\"name\":\"arXiv - PHYS - High Energy Physics - Experiment\",\"volume\":\"40 1\",\"pages\":\"\"},\"PeriodicalIF\":0.0000,\"publicationDate\":\"2024-08-25\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"\",\"citationCount\":\"0\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"arXiv - PHYS - High Energy Physics - Experiment\",\"FirstCategoryId\":\"1085\",\"ListUrlMain\":\"https://doi.org/arxiv-2408.13812\",\"RegionNum\":0,\"RegionCategory\":null,\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"\",\"PubModel\":\"\",\"JCR\":\"\",\"JCRName\":\"\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"arXiv - PHYS - High Energy Physics - Experiment","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/arxiv-2408.13812","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
引用次数: 0
摘要
在对撞机物理实验中,粒子识别(PID),即识别探测器中的带电粒子种类通常是数据分析中最关键的工具之一。近十年来,机器学习技术逐渐成为粒子识别的主流方法之一,与经典算法相比,机器学习技术通常具有更强的识别能力。近年来,量子机器学习(QML)在传统机器学习和量子计算技术之间架起了一座桥梁,为传统机器学习模型提供了进一步改进的潜力。在这项工作中,针对$\mu^{\pm} /\pi^{\pm/pi^\{pm}$的辨别问题,我们开发了一种具有九个量子比特的变分量子分类器(VQC)。利用 IBM 量子模拟器,我们研究了各种编码电路和变分算法,以探索它们的性能。经典优化器能够有效地最小化量子经典混合模型中的损失函数。我们将 VQC 与传统的多层感知神经网络进行了比较,发现它们在相同数据集上的表现相似,这说明未来将量子机器学习应用于对撞机物理实验的数据分析是可行的。
Muon/Pion Identification at BESIII based on Variational Quantum Classifier
In collider physics experiments, particle identification (PID), i. e. the
identification of the charged particle species in the detector is usually one
of the most crucial tools in data analysis. In the past decade, machine
learning techniques have gradually become one of the mainstream methods in PID,
usually providing superior discrimination power compared to classical
algorithms. In recent years, quantum machine learning (QML) has bridged the
traditional machine learning and the quantum computing techniques, providing
further improvement potential for traditional machine learning models. In this
work, targeting at the $\mu^{\pm} /\pi^{\pm}$ discrimination problem at the
BESIII experiment, we developed a variational quantum classifier (VQC) with
nine qubits. Using the IBM quantum simulator, we studied various encoding
circuits and variational ansatzes to explore their performance. Classical
optimizers are able to minimize the loss function in quantum-classical hybrid
models effectively. A comparison of VQC with the traditional multiple layer
perception neural network reveals they perform similarly on the same datasets.
This illustrates the feasibility to apply quantum machine learning to data
analysis in collider physics experiments in the future.