FEATURES OF CONSTRUCTION AND BASIC DIRECTIONS OF DEVELOPMENT OF VIRTUAL DIGITAL ASSISTANTS

Abstract

The article analyzes the main aspects of creating virtual assistants that are part of intelligent computer programs – artificial intelligence systems (AI). The main task of “artificial intelligence” is to ensure effective communication of intelligent robotic systems (including unmanned vehicles) with humans. The basis of the above is in-depth training (systematic machine translation, speech recognition, processing of complex texts in natural languages, computer vision, automation of driving, etc.). This machine learning subsystem can be characterized using neural network models that mimic the brain. Any neural network model learns from large data sets, so it acquires some “skills”, but how it uses them remains for engineers, which ultimately becomes one of the most important problems for many deep learning applications. The reason is that such a model is formal and without an understanding of the logic of its actions. This raises the question: is it possible to increase the level of trust in such systems based on machine learning? Machine learning algorithms are complex mathematical descriptions and procedures and have a growing impact on people's lives. As the decision is increasingly determined by the algorithms, they become less transparent and understandable. Based on the foregoing, the paper considers the issues of the technological component and the algorithms of virtual digital assistants, conducts information modeling based on the conceptual model of the interaction of the virtual assistant with the database, and analyzes the scope and further development of the IT-sphere.