The discovery of equations, laws, or principles that are invariant and causal underpins scientific and technical advancement. Robust model discovery has typically emerged from observing the world and, when possible, performing interventions to falsify models.
Recently, data-driven approaches like classic and deep machine learning enabled equation and causal discovery methods to evolve and accelerate progress in computer science, neuroscience, physics and even philosophy, among many applied fields.
We have just published a new study discussing concepts, methods, and relevant works on causal and equation discovery, outlining current key challenges and promising future lines of research. The work also showcases comprehensive case studies in diverse scientific areas ranging from earth and environmental science to neuroscience.
Our tenet is that discovering fundamental laws and causal relations by observing natural phenomena is revolutionised with the coalescence of observational data and simulations, modern machine learning algorithms and domain knowledge. Exciting times are ahead with many challenges and opportunities to improve our understanding of complex systems.
This study is a collaborative work between eight universities in Europe and the United States (Valencia, Berlin, Tübingen, Jena, Stockholm, New York, and Bournemouth Universities).
Camps-Valls, G., Gerhardus, A., Ninad, U., Varando, G., Martius, G., Balaguer-Ballester, E., Vinuesa, R., Diaz, E., Zanna, L. and Runge, J., 2023. Discovering causal relations and equations from data. Physics Reports, 1044, 1-68 (Impact Factor=30).