Architecture of a knowledge processing system

We start with three basic notions: that knowledge is based on concepts; that events enhance or refute such concepts; and that no concept or its modifications is absolute. Within this framework, we build a processing system that is dynamic based on the knowledge available, aware of current events that are modifying the knowledge base with some finite probability that its result is accurate. In a sense the system is limited by its capacity to store/retrieve information (knowledge), its ability to (intelligently) process information, and its ability to compute the confidence level with which it has generated the previous step(s). We fall back on database facilities for storing/retrieving information, on AI techniques for processing and on basic probability (fuzzy set) theory to numerically compute or at least estimate the accuracy of its discrete steps. Whereas any computer system with a complex software structure can serve as a knowledge machine, we present an architecture which stores/retrieves information; processes, learns, and modifies the information; and finally computes or estimates the confidence level in each step or procedure. These are the macro instructions to the rather elaborate hardware. It processes in two dimensions: the knowledge domain (i.e. generates conclusions) and the numeric domain (i.e. generates confidence level), confirming the earlier stated notion that the conclusion reached so far is not undeniable