The one important principle to building difficult System of Systems is Modularity.

I think that a modular approach and philosophy should be use to building of AI Systems as well. It can be a modular combination of different Neural Networks into Big Hierarchical systems etc. But in current time I don't see an implementation of this in a practice.

What is your opinion regarding to this issues? What is the better way to ahead?

More Vadym Slyusar's questions See All
Similar questions and discussions