Dynamic knowledge base

Neural networks have been used in various fields like control, estimation, simulation, AI, etc. But there are only few applications in the areas of semantics and logics which will be very important in the real usage of neural networks for the coming AI age and so on. Now, some research has been done on AI, semantics and logic, but the presented research is only concerned with network relations on semantics or constructing relational trees for data, etc. We know that it is possible to construct some sort of knowledge base using neural networks. But this is a statistical knowledge base which cannot evolve itself situation by situation. Therefore, a dynamic knowledge base is studied by introducing a “sequence memorizing method” by using backpropagation neural networks. By using this method, each knowledge base chosen will be remembered in a sequential order as they are input and can be reproduced with its order which makes it possible to have knowledge bases that change those properties situation by situation, dynamically