摘要
近年来机器学习领域在硬件和算法上均有长足的发展,并作为一种大数据分析方法获得了广泛的应用.其中,机器学习在物理学方面的应用也逐渐热门起来.本文综述了机器学习的各类模型在相对论重离子碰撞物理研究中的应用,包括碰撞参数预测、核形变参数预测、相变分类、流体演化模拟等.其中涉及的机器学习模型既包括集成学习、主成分分析等经典机器学习模型,也包括卷积神经网络、点云网络等深度学习模型.由于机器学习出色的性能和运算效率,它在相对论重离子碰撞物理中的应用将获得越来越多的关注.
Recently,with rapid hardware and algorithms development,machine learning has been widely used as a significant data analysis method.This article reviews the application of different machine learning algorithms in heavy ion collisions,including impact parameter prediction,nuclear deformation parameter prediction,phase transitions classification,fluid evolution simulation,etc.The machine learning algorithms comprise classic machine learning algorithms,such as ensemble learning and principal component analysis,and deep learning algorithms,such as convolutional neural networks and point cloud networks.Because of the excellent performance and efficiency of machine learning,these applications will receive much attention in our field.
作者
周蒙
罗逸群
宋慧超
ZHOU Meng;LUO YiQun;SONG HuiChao(State Key Laboratory of Nuclear Physics and Technology,School of Physics,Peking University,Beijing 100871,China)
出处
《中国科学：物理学、力学、天文学》
CSCD
北大核心
2022年第5期3-15,共13页
Scientia Sinica Physica,Mechanica & Astronomica
基金
国家自然科学基金(编号:12075007,11675004)资助项目。
