A group of physicists from the University of Chile they want to achieve a rather ambitious goal, to develop a method that can predict extreme events of nature, such as earthquakes or floods.
The result is the product of the intellectual communion of Chilean and French scientists and could support the development of applications towards multiple disciplines “we have valuable elements that make us think that this research paves the way to predict extreme events such as droughts, earthquakes, economic and social crises ”, says Marcel Clerc, an academic from the Department of Physics of the Faculty of Physical Sciences of the University of Chile.
The research, which took almost three years, used artificial intelligence methods to predict catastrophic events in mathematical models. “We show that it is possible to forecast occurrences of unlikely extreme events such as gigantic and catastrophic pulsations using neural networks. By selecting regions of maximum information transfer (entropy), we show that it is possible to obtain higher forecast accuracy using non-local data versus local data, allowing longer warning times. In other words, we can anticipate results that we would not have been able to glimpse before”, adds the PhD in Physics from the University of Nice Sophia Antipolis (France).
The research path
The experimental data were carried out in France (Center for Nanoscience and Nanotechnology of the Université Paris-Saclay). By analyzing the dynamics of light emitted by optically excited semiconductor cavities. Personal computers were used for data analysis and the results were verified in parallel computers at the University of Lille.
Now the team plans to apply these methods to other physical systems and other contexts. “Although we do not plan to use it in the social sciences, it is possible to predict collective behavior such as economic crises,” says Clerc, who is also a researcher at the MIRO Millennium Institute of Optics.
Also participating in the paper were VA Pammi and S. Barbay, both from the Paris-Saclay University (CNRS, Center de Nanosciences et de Nanotechnologies) and S. Coulibaly from the University of Lille (France).
The results were published in the article “Extreme Events Prediction from Nonlocal Partial Information in a Spatiotemporally Chaotic Microcavity Laser,” which appeared in the journal Physical Review Letters.