Mapping of Probabilities (a Theory for the Interpretation of Uncertain Physical Measurements)

Albert Tarantola, University of Paris VI.While the usual way for "assimilating" uncertain observations goes through the use of conditional probabilities, one may try to introduce different concepts, directly borrowed from set theory. It is argued that these new concepts complete the traditional Kolmogorov point of view. The resulting theory seems well adapted for facing real world problems, where observations are only used to "falsify" models (à la Popper), these falsifications defining the transition from a prior distribution of models into a posterior distribution (à la Bayes).

Location: 118 Zumberge Hall

Audiences: Everyone Is Invited

Contact: Masako Okamoto