The Way Children Learn Is Not a Recognition Problem

The way children learn is relatively simple, in a way. It's essentially starting from an internal representation of the world that gets more and more combined with otherinternal representations. And out of this combination, you now, due to this combination, children are capable of tackling the problem of combinatorial complexity. The way things combine in our physical world is pretty much infinite. So how would it be possible to learn from a world that has combinatorialcomplexity with just one single neural network? We don't really get it otherwise we would use that particular method, the biological way of learning stuff.