Self evolving dynamic knowledge base

There have been a lot of studies on artificial intelligence. But still we have a lot of problems left, like analysis of the environment for solving problems, low robustness, lack of information on a target, less matured computer languages, and so forth. Which means that a lot of human interactions are still necessary in order to get proper answers for given problems. Studies on how to reduce the human interactions are performed by using backpropagation neural networks with the idea of dynamic knowledge base which has an ability to evolve itself in order to get the most appropriate knowledge base and performance mechanisms for solving problems given. And as a result, it has been proved that it is possible to have a self evolving dynamic knowledge base which reduces a lot of human interactions which used to need.