-
复杂系统领域的研究已有半个世纪的历史,2021年的诺贝尔物理学奖可视为这一领域的里程碑式事件。这一年,奖项授予了三位在复杂系统研究领域做出了杰出贡献的科学家:真锅淑郎(Syukuro Manabe),克劳斯·哈塞尔曼(Klaus Hasselman)和乔治·帕里西(Giorgio Parisi)。前两位科学家因“对地球气候的物理建模、量化可变性和可靠地预测全球变暖”这一贡献获奖;而帕里西的贡献则在于“发现了从原子到行星尺度的物理系统中无序和涨落之间的相互作用”。这一奖项的颁发,不仅肯定了复杂系统科学在揭示科学基本规律和改善人类福祉方面的重要贡献,更标志着这一领域的研究已经进入一个新的历史阶段,学界和社会可以进一步期待复杂系统科学发展带来的更多科学突破和社会应用。
同样,人工智能(Artificial Intelligence, AI)这个最早可追溯到20世纪50年代的研究领域,也在经历了起初的发展和低谷之后,在21世纪第二个十年里取得了革命性进步。2012年,谷歌公司研究团队开发的深度学习神经网络AlexNet在图像分类竞赛中战胜了传统方法,这一事件引发了对深度学习的广泛兴趣;随后,2016年,DeepMind公司开发的AlphaGo击败了围棋世界冠军,这一里程碑事件标志着AI在复杂策略游戏中的巨大突破。在科学领域,AI技术也展现出了令人瞩目的应用,其中一项备受瞩目的成就是DeepMind团队开发的AlphaFold,其利用深度神经网络技术,成功地解决了蛋白质结构预测这一长期以来困扰科学家的难题。这一突破性的成果在2020年引起了广泛关注,并被认为是AI在生物化学领域的重大突破,为新药物研发、疾病治疗等领域打开了全新的可能性,引发了“AI for Science”的研究热潮。
在这样的背景下,复杂系统研究与AI的结合,预示着一种新的可能性。传统复杂系统的研究范式是寻求基本原理的理论模型,然而在很多复杂问题中无法建立有效的理论体系。基于各类真实复杂系统收集的大规模数据,我们有望借助AI在理论难以触及的领域发展出新的知识获取和知识表达方法。这种以数据为中心、AI驱动的复杂性科学研究新范式,有可能突破领域研究瓶颈,帮助我们更好地理解自然界和社会中的各类复杂现象、更可靠地预测和控制系统的未来行为,应对现实世界中的复杂问题和挑战。除了加速复杂性科学的发展,复杂系统与AI及数据科学的结合,将进一步鼓励不同学科之间的交流和合作,推动新的研究领域的形成和发展。
为了更好地了解这一极具前景和价值的研究方向,从而为进一步研究和发展提供参考,本文对AI驱动的复杂系统研究进行了综述。首先回顾复杂系统及复杂性科学早期的研究,整理当前领域研究存在的挑战性难题,进一步探讨AI助力下的复杂系统研究,讨论复杂系统预测模拟与规律发现研究方面融合AI技术的代表性工作及发展方向。在AI为传统复杂系统研究领域带来变革的同时(即AI for Complex System),我们还注意到复杂系统视角下AI理论及技术的潜在发展方向(即Complex System for AI),并在本文中进行了系统性的阐述与总结。
Advancements in Artificial Intelligence-Driven Complex Systems Research
-
摘要: 作为一个研究对象涵盖基本物质、生命体和社会的跨学科研究领域,复杂系统的研究有助于增进对自然和社会现象的理解和预测,在解决人类面临的复杂问题中具有重要价值。这一领域的早期研究积累了海量的各类真实复杂系统数据,在此基础上发展数据密集型、人工智能方法驱动的复杂性科学研究新范式,将为复杂系统的描述、预测与知识发现提供一条全新的路径。该文对人工智能驱动的复杂系统研究进行前瞻性的综述,探讨人工智能助力下的复杂系统研究发展前沿,并分析基于人工智能方法的领域代表性工作,最后讨论复杂系统视角下人工智能理论及技术的潜在发展方向。Abstract: Spanning across disciplines with research interests in fundamental matter, life forms, and societal dynamics, the study of complex systems plays a pivotal role in deciphering and forecasting natural and social phenomena, thereby confronting intricate problems of human concern. The wealth of diverse real-world complex system data accumulated through early research has paved the way for a novel paradigm in complexity science research, which is intensively data-driven and steered by Artificial Intelligence (AI) methodologies. This innovative approach provides fresh insights into the characterization, forecasting, and knowledge extraction of complex systems. This article offers a visionary review of AI-driven studies in complex systems, highlighting the pioneering developments spearheaded by AI. It further scrutinizes exemplary works in the domain that leverage AI methodologies and concludes by contemplating the prospective evolution of AI theory and techniques under the lens of complex systems.
-
Key words:
- complex system /
- artificial intelligence /
- machine learning /
- data science
-
[1] 宋学锋. 复杂性、复杂系统与复杂性科学[J]. 中国科学基金, 2003, 17(5): 262-269. SONG X F. Complexity, complex system, and the science of complexity[J]. Bulletin of National Science Foundation of China, 2003, 17(5): 262-269. [2] 张嗣瀛. 复杂系统、复杂网络自相似结构的涌现规律[J]. 复杂系统与复杂性科学, 2006, 3(4): 41-51. ZHANG S Y. The law of emergence of self-similar structures in complex systems and complex networks[J]. Complex Systems and Complexity Science, 2006, 3(4): 41-51. [3] 张永安, 白志学. 复杂系统研究的重要工具: 细胞自动机及其应用[J]. 自然杂志, 1998, 20(4): 192-196. ZHANG Y A, BAI Z X. A important tool for complex system research-cellular automata and its application[J]. Nature Journal, 1998, 20(4): 192-196. [4] 钱学森, 于景元, 戴汝为. 一个科学新领域: 开放的复杂巨系统及其方法论[J]. 自然杂志, 1990, 12(1): 3-10. QIAN X S, YU J Y, DAI R W. A new scientific field: Open complex giant systems and their methodology[J]. Chinese Journal of Nature, 1990, 12(1): 3-10. [5] 周涛, 柏文洁, 汪秉宏, 等. 复杂网络研究概述[J]. 物理, 2005, 34(1): 31-36. ZHOU T, BAI W J, WANG B H, et al. A brief review of complex networks[J]. Physics, 2005, 34(1): 31-36. [6] 汪秉宏, 周涛, 王文旭, 等. 当前复杂系统研究的几个方向[J]. 复杂系统与复杂性科学, 2008, 5(4): 21-28. WANG B H, ZHOU T, WANG W X, et al. Several directions in complex system research[J]. Complex Systems and Complexity Science, 2008, 5(4): 21-28. [7] BOCCALETTI S, BIANCONI G, CRIADO R, et al. The structure and dynamics of multilayer networks[J]. Physics Reports, 2014, 544(1): 1-22. doi: 10.1016/j.physrep.2014.07.001 [8] HOLME P, SARAMÄKI J. Temporal networks[J]. Physics Reports, 2012, 519(3): 97-125. doi: 10.1016/j.physrep.2012.03.001 [9] BARTHÉLEMY M. Spatial networks[J]. Physics Reports, 2011, 499(1-3): 1-101. [10] BIANCONI G. Higher-order networks[M]. Cambridge: Cambridge University Press, 2021. [11] LYU L Y, ZHOU T, ZHANG Q M, et al. The H-index of a network node and its relation to degree and coreness[J]. Nature Communications, 2016, 7: 10168. doi: 10.1038/ncomms10168 [12] FAN T L, LYU L Y, SHI D H, et al. Characterizing cycle structure in complex networks[J]. Communications Physics, 2021, 4: 272. doi: 10.1038/s42005-021-00781-3 [13] CALLAWAY D S, NEWMAN M E, STROGATZ S H, et al. Network robustness and fragility: Percolation on random graphs[J]. Physical Review Letters, 2000, 85(25): 5468-5471. doi: 10.1103/PhysRevLett.85.5468 [14] MENCK P J, HEITZIG J, MARWAN N, et al. How basin stability complements the linear-stability paradigm[J]. Nature Physics, 2013, 9: 89-92. doi: 10.1038/nphys2516 [15] GAO J X, BARZEL B, BARABÁSI A L. Universal resilience patterns in complex networks[J]. Nature, 2016, 536(7615): 238. [16] LIU Y Y, SLOTINE J J, BARABÁSI A L. Controllability of complex networks[J]. Nature, 2011, 473(7346): 167-173. doi: 10.1038/nature10011 [17] YAN G, VÉRTES P E, TOWLSON E K, et al. Network control principles predict neuron function in the Caenorhabditis elegans connectome[J]. Nature, 2017, 550(7677): 519-523. doi: 10.1038/nature24056 [18] YAN G, ZHOU T, HU B, et al. Efficient routing on complex networks[J]. Physical Review E, Statistical, Nonlinear, and Soft Matter Physics, 2006, 73(4): 046108. [19] 韩筱璞, 汪秉宏, 周涛. 人类行为动力学研究[J]. 复杂系统与复杂性科学, 2010, 7(S1): 132-144. HAN X P, WANG B H, ZHOU T. Researches of human dynamics[J]. Complex Systems and Complexity Science, 2010, 7(S1): 132-144. [20] XIE J R, MENG F H, SUN J C, et al. Detecting and modelling real percolation and phase transitions of information on social media[J]. Nature Human Behaviour, 2021, 5(9): 1161-1168. doi: 10.1038/s41562-021-01090-z [21] XIE J R, WANG X R, FENG L, et al. Indirect influence in social networks as an induced percolation phenomenon[J]. Proceedings of the National Academy of Sciences of the United States of America, 2022, 119(9): e2100151119. [22] BURY T M, SUJITH R I, PAVITHRAN I, et al. Deep learning for early warning signals of tipping points[J]. Proceedings of the National Academy of Sciences of the United States of America, 2021, 118(39): e2106140118. [23] XU F L, WU L F, EVANS J. Flat teams drive scientific innovation[J]. Proceedings of the National Academy of Sciences of the United States of America, 2022, 119(23): e2200927119. [24] VLACHAS P R, ARAMPATZIS G, UHLER C, et al. Multiscale simulations of complex systems by learning their effective dynamics[J]. Nature Machine Intelligence, 2022, 4: 359-366. doi: 10.1038/s42256-022-00464-w [25] MURPHY C, LAURENCE E, ALLARD A. Deep learning of contagion dynamics on complex networks[J]. Nature Communications, 2021, 12(1): 4720. doi: 10.1038/s41467-021-24732-2 [26] PATHAK J, HUNT B, GIRVAN M, et al. Model-free prediction of large spatiotemporally chaotic systems from data: A reservoir computing approach[J]. Physical Review Letters, 2018, 120(2): 024102. doi: 10.1103/PhysRevLett.120.024102 [27] LYU L Y, ZHOU T. Link prediction in complex networks: A survey[J]. Physica A: Statistical Mechanics and Its Applications, 2011, 390(6): 1150-1170. doi: 10.1016/j.physa.2010.11.027 [28] ZHOU T. Progresses and challenges in link prediction[J]. iScience, 2021, 24(11): 103217. doi: 10.1016/j.isci.2021.103217 [29] MUSCOLONI A, CANNISTRACI C V. “Stealing fire or stacking knowledge” by machine intelligence to model link prediction in complex networks[J]. iScience, 2023, 26(1): 105697. doi: 10.1016/j.isci.2022.105697 [30] LIU C, ZHAN Y B, LI C, et al. Graph pooling for graph neural networks: Progress, challenges, and opportunities[EB/OL]. [2023-09-03]. http://arxiv.org/abs/2204.07321. [31] ALMAGRO P, BOGUÑÁ M, ÁNGELES SERRANO M. Detecting the ultra low dimensionality of real networks[J]. Nature Communications, 2022, 13(1): 6096. doi: 10.1038/s41467-022-33685-z [32] VILLEGAS P, GILI T, CALDARELLI G, et al. Laplacian renormalization group for heterogeneous networks[J]. Nature Physics, 2023, 19: 445-450. doi: 10.1038/s41567-022-01866-8 [33] SUN F, LIU Y, WANG J X, et al. Symbolic physics learner: Discovering governing equations via Monte Carlo tree search[EB/OL]. [2023-09-03]. https://arxiv.org/abs/2205.13134. [34] SHI H, DING J, CAO Y, et al. Learning symbolic models for graph-structured physical mechanism[C]//The Eleventh International Conference on Learning Representations. Kigali: ICLR, 2023. [35] GAO T T, YAN G. Autonomous inference of complex network dynamics from incomplete and noisy data[J]. Nature Computational Science, 2022, 2: 160-168. doi: 10.1038/s43588-022-00217-0 [36] CHEN Z, LIU Y, SUN H. Physics-informed learning of governing equations from scarce data[J]. Nature Communications, 2021, 12(1): 6136. doi: 10.1038/s41467-021-26434-1 [37] WANG H, MA C, CHEN H S, et al. Full reconstruction of simplicial complexes from binary contagion and Ising data[J]. Nature Communications, 2022, 13(1): 3043. doi: 10.1038/s41467-022-30706-9 [38] SANTORO A, BATTISTON F, PETRI G, et al. Higher-order organization of multivariate time series[J]. Nature Physics, 2023, 19: 221-229. [39] PIAGGESI S, PANISSON A, PETRI G. Effective higher-order link prediction and reconstruction from simplicial complex embeddings[EB/OL]. [2023-10-02]. https://proceedings.mlr.press/v198/piaggesizza.html. [40] BLAZEK P J, VENKATESH K, LIN M M. Automated discovery of algorithms from data[J]. Nature Computational Science, 2024, 4: 110-118. doi: 10.1038/s43588-024-00593-9 [41] ROMERA-PAREDES B, BAREKATAIN M, NOVIKOV A, et al. Mathematical discoveries from program search with large language models[J]. Nature, 2024, 625(7995): 468-475. doi: 10.1038/s41586-023-06924-6 [42] BAKHTIN A, BROWN N, NOVIKOW A, et al. Human-level play in the game of Diplomacy by combining language models with strategic reasoning[J]. Science, 2022, 378(6624): 1067-1074. doi: 10.1126/science.ade9097 [43] RAHWAN I, CEBRIAN M, OBRADOVICH N, et al. Machine behaviour[J]. Nature, 2019, 568: 477-486. doi: 10.1038/s41586-019-1138-y [44] MEENA C, HENS C, ACHARYYA S, et al. Emergent stability in complex network dynamics[J]. Nature Physics, 2023, 19: 1033-1042. doi: 10.1038/s41567-023-02020-8 [45] GRZIWOTZ F, CHANG C W, DAKOS V, et al. Anticipating the occurrence and type of critical transitions[J]. Science Advances, 2023, 9(1): eabq4558. doi: 10.1126/sciadv.abq4558 [46] 高庆, 吕金虎. AI促进数学理论研究新范式: 关于复杂系统的一些思考[J]. 中国科学基金, 2022, 36(1): 107-109. GAO Q, LYU J H. New paradigm on mathematical research promoted by AI: Thoughts about complex systems[J]. Bulletin of National Natural Science Foundation of China, 2022, 36(1): 107-109. [47] LI R K, WANG H D, LI Y. Learning slow and fast system dynamics via automatic separation of time scales[C]//Proceedings of the Proceedings of the 29th ACM SIGKDD Conference on Knowledge Discovery and Data Mining. New York: ACM, 2023: 4380–4390. [48] ZHANG G Z, YU Z H, JIN D P, et al. Physics-infused machine learning for crowd simulation[C]//Proceedings of the Proceedings of the 28th ACM SIGKDD Conference on Knowledge Discovery and Data Mining. New York: ACM, 2022: 2439–2449. [49] RU X L, MURDOCH MOORE J, ZHANG X Y, et al. Inferring patient zero on temporal networks via graph neural networks[J]. Proceedings of the AAAI Conference on Artificial Intelligence, 2023, 37(8): 9632-9640. doi: 10.1609/aaai.v37i8.26152 [50] CORNELIO C, DASH S, AUSTEL V, et al. Combining data and theory for derivable scientific discovery with AI-Descartes[J]. Nature Communications, 2023, 14(1): 1777. doi: 10.1038/s41467-023-37236-y [51] 王飞跃. 人工社会、计算实验、平行系统: 关于复杂社会经济系统计算研究的讨论[J]. 复杂系统与复杂性科学, 2004, 1(4): 25-35. WANG F Y. Artificial societies, computational experiments, and parallel systems: A discussion on computational theory of complex social-economic systems[J]. Complex Systems and Complexity Science, 2004, 1(4): 25-35. [52] 王芳, 郭雷. 数字化社会的系统复杂性研究[J]. 管理世界, 2022, 38(9): 208-221. WANG F, GUO L. Research on system complexity of the digital society[J]. Journal of Management World, 2022, 38(9): 208-221. [53] XU F L, LI Y, JIN D P, et al. Emergence of urban growth patterns from human mobility behavior[J]. Nature Computational Science, 2021, 1(12): 791-800. doi: 10.1038/s43588-021-00160-6 [54] CALDARELLI G, ARCAUTE E, BARTHELEMY M, et al. The role of complexity for digital twins of cities[J]. Nature Computational Science, 2023, 3(5): 374-381. doi: 10.1038/s43588-023-00431-4 [55] 李伯虎, 柴旭东, 张霖, 等. 面向新型人工智能系统的建模与仿真技术初步研究[J]. 系统仿真学报, 2018, 30(2): 349-362. LI B H, CHAI X D, ZHANG L, et al. Preliminary study of modeling and simulation technology oriented to neo-type artificial intelligent systems[J]. Journal of System Simulation, 2018, 30(2): 349-362. [56] 苏竣, 魏钰明, 黄萃. 社会实验: 人工智能社会影响研究的新路径[J]. 中国软科学, 2020(9): 132-140. SU J, WEI Y M, HUANG C. Social experiment: A new approach to the study of the social impact of artificial intelligence[J]. China Soft Science, 2020(9): 132-140. [57] BASSETT D S, SPORNS O. Network neuroscience[J]. Nature Neuroscience, 2017, 20(3): 353-364. doi: 10.1038/nn.4502 [58] BARABÁSI DL, BIANCONI G, BULLMORE E, et al. Neuroscience needs network science[EB/OL]. [2023-10-02]. https://arxiv.org/abs/2305.06160. [59] SUN J C, FENG L, XIE J R, et al. Revealing the predictability of intrinsic structure in complex networks[J]. Nature Communications, 2020, 11(1): 574. doi: 10.1038/s41467-020-14418-6 [60] NGUYEN T M, THOMAS L A, RHOADES J L, et al. Structured cerebellar connectivity supports resilient pattern separation[J]. Nature, 2023, 613(7944): 543-549. doi: 10.1038/s41586-022-05471-w [61] SEROUSSI I, NAVEH G, RINGEL Z. Separation of scales and a thermodynamic description of feature learning in some CNNs[J]. Nature Communications, 2023, 14: 908. doi: 10.1038/s41467-023-36361-y [62] DAVIES A, VELIČKOVIĆ P, BUESING L, et al. Advancing mathematics by guiding human intuition with AI[J]. Nature, 2021, 600(7887): 70-74. doi: 10.1038/s41586-021-04086-x [63] PARK J S, O’BRIEN J, CAI C J, et al. Generative agents: Interactive simulacra of human behavior[C]//Proceedings of the 36th Annual ACM Symposium on User Interface Software and Technology. New York: ACM, 2023. [64] WEI J, TAY Y, BOMMASANI R, et al. Emergent Abilities of Large Language Models[EB/OL]. [2023-10-02]. https://arvix.org/abs/2206.
计量
- 文章访问数: 6155
- HTML全文浏览量: 4112
- PDF下载量: 39
- 被引次数: 0