Understanding causation via correlations and linear response theory

by Angelo Vulpiani (Sapienza Università di Roma)

Wednesday 12 May 2021, 14:30 → 16:00 Europe/Rome

In spite of the (correct) common-wisdom statement correlation does not imply causation, a proper employ of time correlations and of fluctuation-response theory allows us to understand causal relations between variables of a multidimensional linear Markov process.  It is shown that fluctuation-response formalism can be used both to find direct causal links between variables of a system and to introduce a degree of causation, cumulative in time, whose physical interpretation is straightforward. Although for generic nonlinear dynamics there is no simple exact relationship between correlations and response functions....

Link per il collegamento: https://meet.google.com/pip-kvzt-fkz