Bar des Sciences’ “Artifical intelligence: Shannon’s legacy”
Shannon succeeded in classifying data, but removed the whole concept of relevance in the process. This is unacceptable for artificial intelligence. A program should be relevant to be viewed as intelligent. Data relevance is the holy grail for AI. How do we move up from the concept of data in order to integrate the notions of interest, importance and relevance to context? This is now possible, in principle and soon in practice. Jean-Louis Dessalles, researcher at Télécom Paris, and Milad Doueihi aim to shed some light onto this fascinating subject.
At 7.30 PM, at café du Pont-Neuf, 14 quai du Louvre, Paris 1er
