Abstract:

The main idea of developing mivars emerged 30 years ago when solving tasks of using both databases and logical inference engines for tackling very large-scale heterogeneous tasks of processing diverse information in real time. From a group of five main processes in computer sciences (informatics) mivars are applied to information accumulation and information processing. The evolutionary mivar data model has been developed in databases, which for the first time has allowed us to implement changeable structures of data representation on the basis of gnoseology concepts Thing, Property, Relation. A new production representation in the form of bipartite mivar network graph has allowed us to solve the task of “Millennium NP = P" and for the first time implement logical inference with the linear complexity. Mivar databases and rules accumulate data as well as allowing us to represent mivar networks in the databases and perform logical processing. Thus, mivars have allowed us to unify databases of logical-computational processing and develop a new powerful mathematical tool for solving tasks spanning the whole range of artificial intelligence field: expert systems, text understanding, image recognition, autonomous intelligent robots.

Keywords: mivar, mivar nets, intelligent systems, artificial intelligence, databases, expert systems, meaning understanding, image recognition, robot, simulation modeling