No 2 (2023)

This study presents a classification and comparative analysis of intelligent system event analysis methods for the detection of multistep cyber-attacks, which are a set of sequential actions of one or more attackers pursuing a specific goal of invasion. The paper studies approaches to multistep cyber-attack detection based on knowledge, such as expert rules and scenarios (sequences) of events. The approaches considered are analyzed according to the following criteria: the method for extracting knowledge about scenarios of system events and attacks, the method for scenario knowledge representation, the method for security events analysis and the security problem to be solved. The paper gives the main advantages and disadvantages of approaches to the multistep cyber-attack detection, as well as possible directions of research in this area.

Parkinson's disease is a neurodegenerative neurological disease which progression can be slowed by accurate and timely diagnosis. In this connection, the development of simple and accessible screening methods is relevant, one of which is the analysis of handwriting and drawing. The paper describes such a method based on the application of fuzzy classifier. The algorithm for formation of fuzzy rules bases, in which mountain clustering is applied after a parameters’ adjustment on concrete data, is offered. The Powell's optimization algorithm is chosen to find parameters. The balanced accuracy and the ratio of the number of rules to the number of training samples is used as the target function. The effectiveness of the proposed algorithm is compared with the classical k-means clustering algorithm and the extreme class feature algorithm.

The article considers the problem of synthesizing the behavior of mobile robotic manipulator when solving tasks in a human-oriented environment. The architecture of the control system is presented, which in an original way integrates the modules responsible localization and mapping, planning the motion of the mobile platform between given points, controlling movement along the planned path, object recognition on sensor data and controlling the manipulator when interacting with recognized objects. Aforementioned components are implemented for an example task of ensuring the mobility of a robotic system in a multifloor office building equipped with elevators. During the experiments, the implemented set of components allowed a real robotic system to use the elevator of an office building.

The article considers problems of theoretical and methodological nature arising in the organization of monitoring of emotional reactions of social networks users by means of automatic text analysis. The difficulties accompanying the monitoring activity at the level of motivation, goal-setting and the choice of a method for determining markers of emotionality are analyzed. The problems of determining the sampling criteria for the study of emotional reactions of users within communities and large groups, taking into account textual and non-textual parameters are studied. Difficulties in interpreting the data of a set of psycholinguistic features based on machine learning and questions about the applicability of the classifier built on the training corpus to texts from different fields are outlined.

The use of parabola and its branches as a nonlinearity expanding the logical capabilities of artificial neurons is considered. In particular, the applicability of parabola branches for constructing an s-shaped function suitable for tuning a neural network by reverse error propagation is determined. The implementation of the XOR function on two and three neurons using the proposed approach is demonstrated. The main advantage of the parabola over the sigmoid is a simpler implementation, which speeds up the work of artificial neural networks.

The Twentieth National Conference on Artificial Intelligence with international participation (CAI-2021) was held in Moscow, Russia, on December 21-23, 2022. The conference was coorganized by the Russian Association of Artificial Intelligence, Federal Research Center “Computer Science and Control” of the Russian Academy of Sciences, National Research University “MPEI”. Conference co-chairs are Academician of the RAS I. A. Sokolov (FRC CSC RAS) and Professor N. D. Rogalev (NRU MPEI). Various areas of artificial intelligence were presented in plenary reports and at section meetings.