Developing a formal calculus for brain computation
Computer science and brain research are collaborating to discover the atomic operations that make the mind emerge from the brain.
I consider Artificial Intelligence (AI) as the Holy Grail of computer science. Having a computer with human-level comprehension would be the killer app of the 21st century. Recently, we made great strides in the field and reached optimal performances in some contexts. We are talking about neural networks capable of discerning faces and recognizing objects in photos. These are complicated tasks that would be difficult to program classically. So why don’t we have sentient robots going around our cities? Well, current artificial intelligence is not that intelligent. I like to call it “statistics on steroids” because that is what it is. These algorithms are usually feature detectors trained on specific tasks. They are powerful, yet their application remains limited to a particular circumstance. In other terms, we would like broader AI models capable of better generalization. And what do we do when we search for elegant and practical solutions to complicated problems? Simple, we ask Mother Nature.
