AI, consciouness, ML

Artificial Intelligence is certainly one of the main ideas that are going to shape the upcoming century. It is a very large problem with tremendous applications and it is intimately related to many very exciting questions like robotics, artificial motion, locomotion, computer vision, and so on. See Barack Obama interview for Wired

Below, I discuss some of the hot topics related to the development of Machine Learning, deep learning, artificial intelligence and co.

Most Machine learning systems as well as algorithms dealing with large data bases are now desperately relying on improvements on chips. Anybody that has run a deep network on a single CPU versus multiple GPUs knows how much of a difference this can make. Clearly thus, machine learning applications require more than CPU power.

Interesting introductions/references are


Tensor flow

The sandpile model

Additional References :


At the heart of the artificial intelligence question lies the philosophical question of consciousness which has so far been the prerogative of philosophers and neurologists. Many great minds have been interested in that topic throughout the years without ever answering the question:

“What is a full characterization of consciousness ?”

It includes Locke, Kant, Schrodinger, Hegel, Freud, Michel Bitbol, Sir Roger Penrose, Stuart Hameroff and more recently Max Tegmark. Those are serious references yet many of the references on the subject are pretences.


Good introductions to the mathematics of machine learning

If you look at how humans learn, it ismostly unsupervised

MIT Review

Yann leCun for facebook