ARTIFICIAL INTELLIGENCE TECHNOLOGIES IN INFORMATION PROCESSING PROCESSES

Verezubova Natalya Afanasyevna, Petrakova Natalya Vasilievna, Petrakov Mikhail Alexandrovich

At the present stage, characterized by global digitalization, there is a gradual introduction of technologies based on artificial intelligence into various processes. Information processing processes for various purposes are no exception. At the same time, the analysis of the use of artificial intelligence in information processing processes has shown that a simple decision on the introduction of artificial intelligence in these processes is not enough. It is necessary to define neural network training algorithms and choose the type of neural network to be trained based on the purposes of information processing, as well as formulate a list of activities performed for the purposes of using neural networks. For example, an incorrectly selected neural network and the specified training data can lead to critical consequences when processing information. This is due to the fact that neural networks, which are the basis of artificial intelligence, evolve on the basis of errors while learning, and such errors will increase with each stage of training. At the same time, errors in the actions of artificial intelligence may not be noticeable for a long time, even with its practical application. Incorrectly selected types of neural networks that do not correspond to the nature of the information being processed can also lead to problems related to the interpretation of information. For example, artificial intelligence can produce a certain result, but this is not always the case, such a result will meet the quality requirements and cannot be applied in practice.

Keywords:artificial intelligence, information processing, neural networks, neural network training, machine learning, learning algorithms