SELF-ORGANIZATION OF MATHEMATICAL MODELS FOR CREATING AN ARTIFICIAL INTELLIGENCE SYSTEM.

True relationships and classifications and long range numerical predictions of random processes can be found from short samples of observations by sorting many candidate variants according to the consistency criterion or the criterion of balance of variables. Here lies the advantage of a computer (programmable according to the principle of self-organization of models) over man for whom, while solving these problems, lengthy verification of relationships in practice is necessary. Computer criteria of truth can serve as a basis for creating artificial intelligence.