{"title":"生成点云时注意平均值域","authors":"Benno Käch, Isabell Melzer-Pellmann, Dirk Krücker","doi":"arxiv-2408.04997","DOIUrl":null,"url":null,"abstract":"Collider data generation with machine learning has become increasingly\npopular in particle physics due to the high computational cost of conventional\nMonte Carlo simulations, particularly for future high-luminosity colliders. We\npropose a generative model for point clouds that employs an attention-based\naggregation while preserving a linear computational complexity with respect to\nthe number of points. The model is trained in an adversarial setup, ensuring\ninput permutation equivariance and invariance for the generator and critic,\nrespectively. To stabilize known unstable adversarial training, a feature\nmatching loss is introduced. We evaluate the performance on two different\ndatasets. The former is the top-quark \\textsc{JetNet150} dataset, where the\nmodel outperforms the current state-of-the-art GAN-based model, despite having\nsignificantly fewer parameters. The latter is dataset 2 of the CaloChallenge,\nwhich comprises point clouds with up to $30\\times$ more points compared to the\nfirst dataset. The model and its corresponding code are available at\n\\url{https://github.com/kaechb/MDMA/tree/NeurIPS}.","PeriodicalId":501181,"journal":{"name":"arXiv - PHYS - High Energy Physics - Experiment","volume":"75 1","pages":""},"PeriodicalIF":0.0000,"publicationDate":"2024-08-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":"{\"title\":\"Pay Attention To Mean Fields For Point Cloud Generation\",\"authors\":\"Benno Käch, Isabell Melzer-Pellmann, Dirk Krücker\",\"doi\":\"arxiv-2408.04997\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"Collider data generation with machine learning has become increasingly\\npopular in particle physics due to the high computational cost of conventional\\nMonte Carlo simulations, particularly for future high-luminosity colliders. We\\npropose a generative model for point clouds that employs an attention-based\\naggregation while preserving a linear computational complexity with respect to\\nthe number of points. The model is trained in an adversarial setup, ensuring\\ninput permutation equivariance and invariance for the generator and critic,\\nrespectively. To stabilize known unstable adversarial training, a feature\\nmatching loss is introduced. We evaluate the performance on two different\\ndatasets. The former is the top-quark \\\\textsc{JetNet150} dataset, where the\\nmodel outperforms the current state-of-the-art GAN-based model, despite having\\nsignificantly fewer parameters. The latter is dataset 2 of the CaloChallenge,\\nwhich comprises point clouds with up to $30\\\\times$ more points compared to the\\nfirst dataset. The model and its corresponding code are available at\\n\\\\url{https://github.com/kaechb/MDMA/tree/NeurIPS}.\",\"PeriodicalId\":501181,\"journal\":{\"name\":\"arXiv - PHYS - High Energy Physics - Experiment\",\"volume\":\"75 1\",\"pages\":\"\"},\"PeriodicalIF\":0.0000,\"publicationDate\":\"2024-08-09\",\"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-2408.04997\",\"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-2408.04997","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
由于传统的蒙特卡洛模拟计算成本高昂,特别是对于未来的高亮度对撞机,利用机器学习生成对撞机数据在粒子物理学中越来越受欢迎。我们提出了一种点云生成模型,它采用了基于注意力的聚合,同时保持了与点数有关的线性计算复杂性。该模型在对抗设置中进行训练,分别确保生成器和批判者的输入包络相等性和不变性。为了稳定已知的不稳定对抗训练,我们引入了特征匹配损失。我们对两个不同数据集的性能进行了评估。前者是顶夸克文本集(top-quark textsc{JetNet150} dataset),在该数据集上,尽管模型的参数明显较少,但其性能却优于目前最先进的基于 GAN 的模型。后者是 CaloChallenge 的数据集 2,与第一个数据集相比,该数据集包含的点云多达 30 倍。该模型及其相应代码可在以下网址获取:url{https://github.com/kaechb/MDMA/tree/NeurIPS}。
Pay Attention To Mean Fields For Point Cloud Generation
Collider data generation with machine learning has become increasingly
popular in particle physics due to the high computational cost of conventional
Monte Carlo simulations, particularly for future high-luminosity colliders. We
propose a generative model for point clouds that employs an attention-based
aggregation while preserving a linear computational complexity with respect to
the number of points. The model is trained in an adversarial setup, ensuring
input permutation equivariance and invariance for the generator and critic,
respectively. To stabilize known unstable adversarial training, a feature
matching loss is introduced. We evaluate the performance on two different
datasets. The former is the top-quark \textsc{JetNet150} dataset, where the
model outperforms the current state-of-the-art GAN-based model, despite having
significantly fewer parameters. The latter is dataset 2 of the CaloChallenge,
which comprises point clouds with up to $30\times$ more points compared to the
first dataset. The model and its corresponding code are available at
\url{https://github.com/kaechb/MDMA/tree/NeurIPS}.