Discovering causal relations and equations from data

被引:31
作者
Camps-Valls, Gustau [1 ,11 ]
Gerhardus, Andreas [2 ]
Ninad, Urmi [2 ,3 ]
Varando, Gherardo [1 ]
Martius, Georg [4 ,5 ]
Balaguer-Ballester, Emili [6 ,7 ,8 ]
Vinuesa, Ricardo [9 ]
Diaz, Emiliano [1 ]
Zanna, Laure [10 ]
Runge, Jakob [2 ,3 ]
机构
[1] Univ Valencia, Valencia, Spain
[2] German Aerosp Ctr, Jena, Germany
[3] Tech Univ Berlin, Berlin, Germany
[4] Univ Tubingen, Tubingen, Germany
[5] Max Planck Inst Intelligent Syst, Tubingen, Germany
[6] Bournemouth Univ, Bournemouth, England
[7] Med Fac Mannheim, Mannheim, Germany
[8] Heidelberg Univ, Mannheim, Germany
[9] KTH Royal Inst Technol, FLOW, Engn Mech, Stockholm, Sweden
[10] NYU, New York, NY USA
[11] Univ Valencia, Image Proc Lab IPL, E4 Bldg 4th Floor,Parc Cientif,C Cat Agustin Escar, Paterna 46980, Valencia, Spain
来源
PHYSICS REPORTS-REVIEW SECTION OF PHYSICS LETTERS | 2023年 / 1044卷
基金
欧洲研究理事会; 欧盟地平线“2020”; 美国国家科学基金会;
关键词
Causal inference; Causal discovery; Complex systems; Nonlinear dynamics; Equation discovery; Knowledge discovery; Understanding; Artificial intelligence; Neuroscience; Climate science; NONLINEAR GRANGER CAUSALITY; MODEL-REDUCTION; SOIL-MOISTURE; TIME-SERIES; CONDITIONAL-INDEPENDENCE; SCIENTIFIC DISCOVERY; COMPLEX NETWORKS; GRAPHICAL MODELS; PATTERN-ANALYSIS; INFERENCE;
D O I
10.1016/j.physrep.2023.10.005
中图分类号
O4 [物理学];
学科分类号
0702 ;
摘要
Physics is a field of science that has traditionally used the scientific method to answer questions about why natural phenomena occur and to make testable models that explain the phenomena. Discovering equations, laws, and principles that are invariant, robust, and causal has been fundamental in physical sciences throughout the centuries. Discoveries emerge from observing the world and, when possible, performing interventions on the system under study. With the advent of big data and data-driven methods, the fields of causal and equation discovery have developed and accelerated progress in computer science, physics, statistics, philosophy, and many applied fields. This paper reviews the concepts, methods, and relevant works on causal and equation discovery in the broad field of physics and outlines the most important challenges and promising future lines of research. We also provide a taxonomy for data-driven causal and equation discovery, point out connections, and showcase comprehensive case studies in Earth and climate sciences, fluid dynamics and mechanics, and the neurosciences. This review demonstrates that discovering fundamental laws and causal relations by observing natural phenomena is revolutionised with the efficient exploitation of observational data and simulations, modern machine learning algorithms and the combination with domain knowledge. Exciting times are ahead with many challenges and opportunities to improve our understanding of complex systems.(c) 2023 The Author(s). Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).
引用
收藏
页码:1 / 68
页数:68
相关论文
共 50 条
  • [31] Enhancing Causal Discovery from Robot Sensor Data in Dynamic Scenarios
    Castri, Luca
    Mghames, Sariah
    Hanheide, Marc
    Bellotto, Nicola
    CONFERENCE ON CAUSAL LEARNING AND REASONING, VOL 213, 2023, 213 : 243 - 258
  • [32] Learning Causal Effects From Observational Data in Healthcare: A Review and Summary
    Shi, Jingpu
    Norgeot, Beau
    FRONTIERS IN MEDICINE, 2022, 9
  • [33] Causal Relationships in Longitudinal Observational Data: An Integrative Modeling Approach
    Biazoli Jr, Claudinei E.
    Sato, Joao R.
    Pluess, Michael
    PSYCHOLOGICAL METHODS, 2024, 29 (06) : 1132 - 1147
  • [34] Discovering interpretable Lagrangian of dynamical systems from data
    Tripura, Tapas
    Chakraborty, Souvik
    COMPUTER PHYSICS COMMUNICATIONS, 2024, 294
  • [35] Discovering governing equations from partial measurements with deep delay autoencoders
    Bakarji, Joseph
    Champion, Kathleen
    Nathan Kutz, J.
    Brunton, Steven L.
    PROCEEDINGS OF THE ROYAL SOCIETY A-MATHEMATICAL PHYSICAL AND ENGINEERING SCIENCES, 2023, 479 (2276):
  • [36] Discovering causal graphs with cycles and latent confounders: An exact branch-and-bound approach
    Rantanen, Kari
    Hyttinen, Antti
    Jarvisalo, Matti
    INTERNATIONAL JOURNAL OF APPROXIMATE REASONING, 2020, 117 (29-49) : 29 - 49
  • [37] LEARNING DETERMINISTIC CAUSAL RELATIONS
    Feng, Dingcheng
    Chen, Feng
    Xu, Wenli
    INTERNATIONAL JOURNAL ON ARTIFICIAL INTELLIGENCE TOOLS, 2013, 22 (03)
  • [38] The estimation of causal effects from observational data
    Winship, C
    Morgan, SL
    ANNUAL REVIEW OF SOCIOLOGY, 1999, 25 : 659 - 706
  • [39] Objective causal predictions from observational data
    Cox Jr, Louis Anthony
    CRITICAL REVIEWS IN TOXICOLOGY, 2024,
  • [40] Causal Inference From Observational Data: It Is Complicated
    Shpitser, Ilya
    Kudchadkar, Sapna R.
    Fackler, James
    PEDIATRIC CRITICAL CARE MEDICINE, 2021, 22 (12) : 1093 - 1096