Probabilistic inference in Physics illustrated with a toy model

by Prof. Giulio D'Agostini (ROMA1)

Thursday 28 Apr 2016, 14:00 → 16:00 Europe/Rome

Sala Riunioni (Dipartimento di Fisica e Astronomia)

The aim of the seminar is to introduce, making use of a very simple although instructive 'experiment', the probabilistic reasoning to solve what Poincaré called ``the essential problem of the experimental method'', i.e. how to evaluate the probabilities of the possible causes that might have generated observed effects. The approach illustrated is mainly that developed by Laplace, although it is nowadays better known with the appellative 'Bayesian'. Particular attention will be put in the graphical representation of inferential and predictive issues ('Bayesian network')

Slides Bologna_28apr2016.pdf