统计物理和层展论的胜利:为何2024年诺贝尔物理学奖授予神经网络研究  

The triumph of statistical physics andemergentism

作　　者：施郁 SHI Yu(CAS Center for Excellence in Quantum Information and Quantum Physics,University of Science and Technology of China,Shanghai 201315;University of Science and Technology of China,Hefei 230026,China;Department of Physics,Fudan University,Shanghai 200438)

机构地区：[1]中国科学院量子信息与量子科技创新研究院和中国科学技术大学上海研究院,上海201315 [2]中国科学技术大学,安徽合肥230026 [3]复旦大学物理学系,上海200438

出　　处：《大学物理》2025年第1期1-16,33,共17页College Physics

基　　金：国家自然科学基金(T2241005,12447214)资助。

摘　　要：本文详细介绍2024年诺贝尔物理学奖获奖成果的来龙去脉,阐述多个相关领域的科学背景、历史、现状和未来,强调神经网络研究的物理学意义和生物学源泉,展示鲜为人知的关联,包括学科之间的关联和科学家之间的关联,反映了学科发展的多维度因素.本文指出,获奖成果标志了理解智能的物理学成就,是统计物理和层展论的主动实践和胜利.物理学未来将以各种新的方式更广泛地存在.We introduce the achievements for which 2024 Nobel Prize in Physics was awarded,and describe the scientific background,history,current status and future,emphasizing the physical implications and biological sources,exposing the connections that were less known,including those among different disciplines and among scientists,indicating many different factors in the development of sciences.It is noted that the achievements winning this prize characterize the success of physics in understanding intelligence,and were initiative practice and triumph of statistical physics and emergentism.The future of physics will exist in various new forms.

关 键 词：诺贝尔物理学奖 神经网络 机器学习 统计物理 层展论 

分 类 号：TP183[自动化与计算机技术—控制理论与控制工程] O4-19[自动化与计算机技术—控制科学与工程]