Probabilistic Networks for Modeling Complex Systems

During the last decades, main methods and tools to study and characterise complex systems have been developed in the area of complex networks in statistical physics. In parallel, Probabilistic Networks (PNs) have been developed as generic modelling and prediction data science methods. PNs exploit conditional independencies in the dataset and compactly represent the joint probability density function P(X)  of a problem domain X with the minimum amount of parameters in a network structure.

This talk explores the methodological synergies between complex networks and probabilistic networks, focusing on the development of suitable approaches to model and predict the behaviour of complex systems from data. The talk will be illustrated at the hand of data of Earth’s climate system.

Presential in the seminar room. Zoom stream: 

https://us06web.zoom.us/j/98286706234?pwd=bm1JUFVYcTJkaVl1VU55L0FiWDRIUT09