INVESTIGATION OF THE HEMPEL PARADOX WITH CLASSES IN BIG DATA OF SOCIO-ECONOMIC SYSTEMS

Abstract

The methods of self-learning and learning with an artificial intelligence teacher are being investigated. The research is carried out in the interests of solving the problems of countering money laundering and terrorist financing on incomplete, heterogeneous big data of socio-economic systems. The N-scheme of the Markov algorithm with a quantitative measure of induction is synthesized. A method is proposed for recording the number of relationships (interactions) of big data objects in the database of artificial intelligence. The theoretical provisions of Hempel’s paradox with classes are considered. The Hempel paradox is evaluated in heterogeneous big data. The Markov occurrence diagram for Hempel’s paradox with classes is synthesized. The Hempel paradox with classes is investigated under conditions of incompleteness and heterogeneity of the data sample. A method of inductive self-learning of artificial intelligence in the N-scheme of the Markov algorithm with a quantitative measure of induction is proposed. The method of using the associator m for training artificial intelligence in the N-scheme of the Markov algorithm is presented. A method of category, predetermined, inductive, teacher-directed self-study is proposed. It is concluded that self-learning methods have a limited potential for mathematical expressiveness. The reason for the low expressiveness is the inability to correctly operate with classes and categories, as well as the limitations of the possibility of logical predestination of truth, taking into account the heterogeneous, anthropogenic nature of objects and morphisms. The study of methods of teaching without a teacher revealed the scientific task of overcoming the Hempel paradox for classes of objects. It is indicated that the proposed method of categorically predetermined, inductive, teacher-directed self-learning actually shows that artificial intelligence a priori must be trained by a teacher, in other words, it cannot correctly self-study without it.