Dynamic knowledge processing - Practical approach

Recently there was a new wave of interest to the real-time AI systems as the ideas of knowledge accumulation and knowledge evolution from data got their full recognition. Our paper seeks a solution through the concept of Dynamic Expert Systems. This concept was proposed by the author some time ago, and a quasi-static approach was developed to implement such a system. The latter permits a smooth combination of static and change, when during a certain amount of time the knowledge used is considered to be stable, yet over a larger time intervals some drastic changes of knowledge are allowed. The resulting Dynamic Expert System never stops, occasionally interrogating the user if it suspects that some of the previously entered data are obsolete. In this aspect the computer system behaves as a "living creature." The proposed approach resolves the classic Frame Problem of Artificial Intelligence. Yet some more sophisticated questions of AI related to the interaction of data and knowledge, are considered in the paper as well. The paper also illustrates a close connection of Dynamic System approach to the area of Empirical Bayesian schemes and to the use of Dynamic Logic in modern Neural Nets proposed by some researches.