Yann LeCun on a vision to make AI systems learn and reason like animals and humans
Intéressant la vision long terme de Yann LeCunet cette modélisation du cerveau

LeCun’s vision requires much deeper exploration than is possible in a brief blog post, and many difficult challenges lie ahead. One of the most interesting and difficult of these is instantiating the details of the architectures and training procedures for the world model. In fact, it could be argued that training world models constitutes the main challenge toward real progress in AI over the next decades.
via Chrome : lire l’article source