Sublime

But the networks themselves were still severely limited in what they could do. Accurate results to complex problems required many layers of artificial neurons, but researchers hadn’t found a way to efficiently train those layers as they were added. Deep learning’s big technical break finally arrived in the mid-2000s, when leading researcher Geoffre

“neocognitron,”

Depuis, il a théorisé son idée de l’hybridation des intelligences humaine et artificielle en utilisant l’image du centaure, produit de ces deux intelligences. Je le rejoins pleinement : l’intelligence collective, dès lors qu’on sait la catalyser et la canaliser – y compris, nous venons de le voir, avec des très jeunes –, renforcée par la puissance

‘Learning representations by back-propagating errors’, written by two cognitive scientists from the University of California San Diego, David Rumelhart and Ronald Williams, and a computer scientist from Carnegie Mellon, Geoffrey Hinton,