Physics and deep learning：an introduction to the 2024 Nobel Prize in Physics
-
摘要: 2024年诺贝尔物理学奖授予神经网络相关的研究工作，充分肯定了以人工神经网络为代表的深度学习方法在多学科交叉前沿中的变革性影响。物理学家约翰·霍普菲尔德与“AI教父”杰弗里·辛顿因其在人工神经网络发展史上的杰出贡献荣膺此奖，引发了学术界的广泛关注与深入讨论。文章将从物理学研究者的视角，解读两位诺奖得主的代表性研究成果，探讨物理学与深度学习的紧密联系，分析物理学在推动深度学习发展中的启示性作用。并以深度学习与第一性原理计算方法的结合为例，展望深度学习对物理学未来发展的深远影响。Abstract: The 2024 Nobel Prize in Physics was awarded for pioneering research on neural networks, recognizing the transformative impact of deep learning across interdisciplinary fields. Physicist John Hopfield and“Godfather of AI”Geoffrey Hinton were honored for their outstanding contributions to the development of artificial neural networks. This article, written from the perspective of physics researchers, will highlight the representative research achievements of the laureates, and explore the deep connection between physics and deep learning. The essential role of physics in advancing deep learning will be examined, and the profound impact of deep learning on the future development of physics will be envisioned, using its integration with first-principles calculations as a concrete example.
-
-
[1] LeCun Y，Bengio Y，Hinton G. Nature，2015，521：436
[2] McCulloch W S，Pitts W. Bull. Math. Biophys.，1943，5：115
[3] Rosenblatt F. Psycholog. Rev.，1958，65：386
[4] Hubel D H，Wiesel T N. J. Physiol.，1959，148：574
[5] Hopfield J J. Proc. Natl. Acad. Sci. USA，1982，79：2554
[6] Ackley D H，Hinton G E，Sejnowski T J. Cogn. Sci.，1985，9：147
[7] Amit D J，Gutfreund H，Sompolinsky H. Phys. Rev. A，1985，32： 1007
[8] Popular Science Background. the Nobel Prize in Physics 2024. https://www.nobelprize.org/uploads/2024/11/popular-physicsprize2024-3.pdf
[9] Hebb D. The Organization of Behavior. A Neuropsychological Theory. Wiley，1949
[10] Amari S I. IEEE Transac. Comput.，1972，100：1197
[11] Hopfield J J，Tank D W. Biol. Cybern.，1985，52：141
[12] Ramsauer H，Schäfl B，Lehner J et al. 2020，arXiv：2008.02217
[13] Nowlan S，Hinton G E. Evaluation of Adaptive Mixtures of Competing Experts. In：Adv. Neural Inf. Process. Syst.3，1990
[14] Srivastava N，Hinton G，Krizhevsky A et al. J. Mach. Learn. Res.，2014，15：1929
[15] Hinton G E，Salakhutdinov R R. Science，2006，313：504
[16] Krizhevsky A，Sutskever I，Hinton G E. Adv. Neural Inf. Process. Syst.，2012，25：1106
[17] Hinton G E. Neural Comput.，2002，14：1771
[18] Tierney L. Ann. Statist.，1994，22：1701
[19] Salakhutdinov R，Hinton G. Deep Boltzmann Machines. In：Artificial intelligence and statistics. PMLR，2009. p.448
[20] Jaitly N，Hinton G. Learning a Better Representation of Speech Soundwaves using Restricted Boltzmann Machines. In：2011 IEEE International Conference on Acoustics，Speech and Signal Processing (ICASSP). IEEE，2011. p.5884
[21] Carleo G，Troyer M. Science，2017，355：602
[22] Silver D，Huang A，Maddison C J et al. Nature，2016，529：484
[23] Jumper J，Evans R，Pritzel A et al. Nature，2021，596：583
[24] Abramson J，Adler J，Dunger J et al. Nature，2024，1：
[25] Udrescu S M，Tegmark M. Sci. Adv.，2020，6：eaay2631
[26] Burger B，Maffettone P M，Gusev V V et al. Nature，2020，583： 237
[27] Martin R M. Electronic Structure：Basic Theory and Practical Methods. Cambridge university press，2020
[28] Kohn W，Sham L J. Phys. Rev.，1965，140：A1133
[29] Kohn W. Phys. Rev. Lett.，1996，76：3168
[30] Das S，Kanungo B，Subramanian V et al. Large-scale Materials Modeling at Quantum Accuracy：Ab Initio Simulations of Quasicrystals and Interacting Extended Defects in Metallic Alloys. In： Proceedings of the International Conference for High Performance Computing，Networking，Storage and Analysis. 2023. p.1
[31] Stocks R，Vallejo J L G，Yu F C et al. Breaking the MillionElectron and 1 EFLOP/s Barriers：Biomolecular-Scale Ab Initio Molecular Dynamics Using MP2 Potentials. In：Proceedings of the International Conference for High Performance Computing， Networking，Storage，and Analysis. 2024. p.1
[32] Perdew J P，Schmidt K. Jacob’s Ladder of Density Functional Approximations for the Exchange-correlation Energy. In：AIP Conference Proceedings. American Institute of Physics，2001. p.1
[33] Behler J，Parrinello M. Phys. Rev. Lett.，2007，98：146401
[34] Zhang L，Han J，Wang H et al. Phys. Rev. Lett.，2018，120： 143001
[35] Batzner S，Musaelian A，Sun L et al. Nat. Commun.，2022，13： 2453
[36] Liao Y L，Wood B M，Das A et al. EquiformerV2：Improved Equivariant Transformer for Scaling to Higher-Degree Representations. In：the Twelfth International Conference on Learning Representations
[37] Merchant A，Batzner S，Schoenholz S S et al. Nature，2023，624：80
[38] Barroso-Luque L，Shuaibi M，Fu X et al. 2024 ，arXiv ：2410. 12771
[39] Li H，Wang Z，Zou N et al. Nat. Comput. Sci.，2022，2：367
[40] Li H，Wang Z，Zou N L et al. 2021，arXiv：2104.03786
[41] Xie T，Grossman J C. Phys. Rev. Lett.，2018，120：145301
[42] Geiger M，Smidt T. 2022，arXiv：2207.09453
[43] Gong X，Li H，Zou N et al. Nat. Commun.，2023，14：2848
[44] Li H，Tang Z，Gong X et al. Nat. Comput. Sci.，2023，3：321
[45] Tang Z，Li H，Lin P et al. Nat. Commun.，2024，15：8815
[46] Li H，Tang Z，Fu J et al. Phys. Rev. Lett.，2024，132：096401
[47] Mortazavi B. Adv. Energy Mater.，2024，2403876：
[48] Unke O，Bogojeski M，Gastegger M et al. Adv. Neural Inf. Processing Syst.，2021，34：14434
[49] Zhong Y，Yu H，Su M et al. npj Comput. Mater.，2023，9：182
[50] Yu H，Xu Z，Qian X et al. Efficient and Equivariant Graph Networks for Predicting Quantum Hamiltonian. In：International Conference on Machine Learning. PMLR，2023. p.40412
[51] Wang Y，Li Y，Tang Z et al. Sci. Bull.，2024，69：30
计量
- 文章访问数: 613
- HTML全文浏览量: 119
- PDF下载量: 1642
下载:
