Context

Complex systems are ubiquitous: from the Earth’s global climate and the human brain, to society and economic systems. The analysis and characterization of complex systems is very challenging: many processes are involved, exhibiting highly nonlinear, dynamic, and networked relations. Different physical, chemical and biological processes interact in several spheres, and at diverse spatio-temporal scales. When data and domain knowledge are available, exploiting regularities through statistical machine learning is certainly an outstanding opportunity.

Our ambition is to develop novel artificial intelligence methods for modeling and understanding complex systems, with particular focus on the Earth, Climate, Brain and Society systems, in isolation and interaction. We want to infer how the systems work from observational data analysis, models, and domain knowledge. A perfect storm is over us: (i) an ever increasing amount of observational data, acquired by all types of sensory systems (from fMRI and biosignal records, to satellite sensors and social network tracking devices), (ii) improved Earth-Climate-Brain mechanistic models capable of resolving processes at higher resolution, accuracy and detail, and (iii) advanced machine learning techniques able to extract patterns and identify drivers from data. In the last decade, machine learning models have helped to analyze and characterize complex systems. ML has helped to monitor and better understand land, oceans, and atmosphere along with their interactions, to explain the emergence of connections and specific behaviors in the brain, and to explain societal and economical relations and impacts.

Why us? Why with us?

Our research agenda for the 2020-2030 follows the tradition and expertise of the team over the last 20 years on remote sensing data analysis and visual neuroscience, and tackles more ambitious problems around complex systems with similar computational techniques and advanced machine learning.