Nº 1 (2023)

Currently, such concepts as “weak” and “strong” artificial intelligence are being frequently applied, but their generally accepted definitions are still missing. In this paper formal and informal suggestions about the essence of artificial intelligence are analyzed. An approach is proposed for quantitative measurement of the "power" of artificial intelligence, which makes it possible to compare various intelligent computer systems with each other.

The paper describes a cloud shell for creating risk assessment systems and predicting the patient's condition based on information from an electronic medical record or other document. The shell integrates various methods and approaches for solving such problems, providing a means of declarative description of the rules for interpreting trained predictive models and knowledge about the dynamics of disease development to generate a detailed explanation. The shell allows you to "collect" in the service for a group of diseases or a section of medicine of interest those implementations of methods for assessing risks and predicting conditions and those knowledge bases about the pathogenesis of diseases that doctors are ready to trust.

The article is about issues of risk reduction when using software solutions based on machine learning methods for classifying chest x-rays on the example of chest x-ray in the diagnosis of bronchopulmonary diseases. A problem statement is formulated to reduce the risk of misdiagnosis by using of methods to counter malicious attacks. The machine learning methods of classification problem, the most dangerous attacks that reduce the recognition efficiency, and measures to counter attacks to reduce risks are identified based on experimental data. These methods were used when experimental studies. Defensive distillation, filtration, unlearning, pruning were used as countermeasures. The results obtained allow us to state that these methods can be used for other images as well. The results of experimental studies made it possible to formulate recommendations as rules, including combinations of recognition methods, attacks, and countermeasures to reduce the risk of misdiagnosis.

This work includes an analytical review, investigated, supplemented and tested actual neural network methods, algorithms and approaches for solving the problem of early detection of fires in forests using images and video streams from unmanned aerial vehicles. The proposed scheme for solving the problem is based on feature extraction and the use of machine learning for frame classification, selection of a rectangular region with target fire sources and accurate semantic segmentation of fires using convolutional neural networks. The performed modifications of the architectures of neural networks are described, which made it possible to improve the F1-measures achieved by them by 20%.

The agenda of the information age requests development of a metrologically sound theory of meaning, reflecting its real nature in human life. This work aims to meet the challenge. First, the paper analyzes premises of classical and applied semiotics, preventing its mathematical formalization. The most unfortunate of them is objectification of meaning, implying the possibility for its modeling based on set calculus. This approach is shown to contradict the pragmatic, creative and subjectively-contextual nature of natural cognition. After Bateson's famous dictum, the problem is solved by grounding meaning in the quantum of subjective behavior - the simplest binary decision. Fragments of the corresponding semantic structure are identified in basic models of emotion, cognitive semantics, functional semiotics, and quantum information science. Alignment of these fragments is shown to reproduce the qubit model of meaningful decision-making based on quantum theory. Integrative potential of this model allows interaction of psychology, cybernetics, behavioral modeling, artificial intelligence, and quantitative semiotics.

Currently, intelligent solutions and artificial intelligence products are being intensively developed for various areas of life, including healthcare. Process of creating and implementing medical AI products is a time-consuming and costly process. The authors of the article consider the potential possibility of accelerating the development and implementation of medical AI products, primarily due to a new solution - the synthetic datasets. The key factors associated with the training datasets collecting are analyzed, including synthetic ones that shorten the development time and improve the quality of products AI based technology.