Grups de recerca
17
des
Seminari FMC: Dimitri Marinelli
El proper dimarts 17 de desembre tindrem el seminari del company Dimitri Marinelli, membre del departament. Us esperem.
Dates:
17-12-2024
Horari:
12:45
Lloc:
Seminar room 320, 3rd floor of the Faculty of Physics, UB
Afegeix-ho a l'agenda (iCal)
Machine Learning for and from Complex Systems
Dimitri Marinelli marinelli@ub.edu
Departament de Física de la Matèria Condensada and UBICS - Institute of Complex Systems. Universitat de Barcelona
The physics of a complex system is determined by the interaction of multiple entities that collectively generate emergent phenomena. Machine learning can help research in this field in at least two ways: (1) by improving our understanding of the physical models, and (2) by introducing new questions that were previously inaccessible through simulations alone, were statistically too intricate to address or based on data otherwise impossible to analyze or gather. This is especially pertinent for analyzing mesoscale effects that extend beyond the main emergent phenomena to investigate processes occurring at smaller scales, potentially down to individual components or agents within the system. On the other hand, complex systems studies have been the origin of research in the field of Artificial Intelligence, and till now they have been helping to improve and understand machine learning models.
In this talk, I will present examples from my research, ranging from complex networks to sociophysics models.
