Automated expert support complex based on a machine learning semantic processor

Abstract

The article is devoted to the creation of an intelligent software complex of a self-learning automated expert system for registering, accounting and executing incidents based on artificial intelligence. The creation of the software complex includes the development of an advisory support expert system based on a semantic processor for speech recognition and synthesis, which allows implementing the first line advisory support based on artificial intelligence. The obtained expert system consists of the following blocks: An input data flow module-GSM (voice address); a semantic processor-the system core in which, using special designed algorithms, the input data flow is converted into the language understood by the system for subsequent intelligent processing set by the specified skills, based on self-learning algorithms; a semantic processor language module with a protocol of recognition, analysis and extraction of input and output data context; an incident registration module with the accounting system in the Redmine platform; PostgreSQL skills database; a manual learning module and an output data flow module. The developed expert support digital ecosystem will allow to completely automate the receipt, processing, registration, accounting and decision-making of incoming requests. The complex implementation can replace first line support call centers and will have a significant economic effect on retail, banking, and consulting systems.