Back to Home page

Brain-like Artificial Intelligence

The new brain concept is a perfecr framework for creating brain-like artificial intelligence. The concept shows what has to be achieved in artificial controlling systems capable of functioning like the brain. The task, despite all its complexity, can be formulated with one sentence: "Build a hierachical system where each level is optimal and can learn, possesses an internal model of its controlled object, uses network computational principle, and you will have brain-like artificial intelligence."

It means that dominating in neurocomputing AI approaches based on detailed simulations of structure and dynamics of selected neural circuits cannot be successful. This is a very simplistic understanding of reverse engineering of the brain. It has all the disadvantages of the classical brain concept.

Instead, AI approach based on the new brain concept should focus on reverse engineering of functional organization of the brain. Any brain-like AI system should be hierarchical to achieve a desired level of complexity, and use neural network computational principle to be reliable and multifunctional. It is important to remember that mechanisms of malfunctions of brain-like AI systems will be similar to those of actual brains.