Alexander Alfimtsev

Verified

Alexander verified their affiliation via an institutional email.

Learn more

Verified

Alexander verified their affiliation via an institutional email.

Learn more

• Doctor of Science (D.Sc.)
• Professor (Full) at Bauman Moscow State Technical University

The integration of artificial intelligence in computer vision tasks has significantly transformed computing technology, especially in applications such as autonomous vehicles, medical image diagnosis, and video surveillance systems. However, multi-object tracking, a key component of modern computer vision, faces challenges such as object occlusion,...

Telecommunications equipment suppliers play an important role in the functioning of corporations with large-scale distributed computer networks such as system integrators, Internet service providers and others. An important condition of sustainable development for these companies is effective and reasonable decision making concerning the choice of...

An analytical review of the models and risks in the researcher’s reproduction system in the scientific specialty “1.2.1. Artificial Intelligence and Machine Learning” is presented. The issues of graduate school management and regulatory barriers in the training of young scientists are considered. Successful practices for defending a PhD thesis at l...

Today, reinforcement learning is one of the most effective machine learning approaches in the tasks of automatically adapting computer systems to user needs. However, implementing this technology into a digital product requires addressing a key challenge: determining the reward model in the digital environment. This paper proposes a usability rewar...

This paper discusses the identification and resolution of social dilemmas in multi-agent reinforcement learning. A multi-agent environment was developed to simulate the movement of autonomous cars on a circular two-lane road, with two agent cars independently trained by a neural network. An independent Q-learning algorithm (IQN) was used for effect...

One of the great challenges in reinforcement learning is learning an optimal behavior in environments with sparse rewards. Solving tasks in such setting require effective exploration methods that are often based on intrinsic rewards. Plenty of real-world problems involve sparse rewards and many of them are further complicated by multi-agent setting...

Creative application of artificial intelligence technologies for content generation is becoming increasingly relevant within the current digital environment. This paper presents an original method of interactive content creation - real-time neurocomics generation based on computer vision data. In the proposed method, the video stream is processed u...

In the rapidly evolving field of reinforcement learning, combination of hierarchical and multiagent learning methods presents unique challenges and opens up new opportunities. This paper discusses a combination of multilevel hierarchical learning with subgoal discovery and multiagent reinforcement learning with hindsight experience replay. Combinin...

In the conditions of rapidly changing market environment and an increase of volume of information that requires analysis and processing before making decisions, the task of creating effective decision support systems is becoming increasingly urgent. These systems allow to reduce the time required for data analysis and development of optimal solutio...

The modern tourist industry is characterized by an abundance of applied multicriteria decision-making tasks. Several researchers have demonstrated that such tasks can be effectively resolved using aggregation operators based on fuzzy integrals and fuzzy measures. At the same time, the implementation of this mathematical tool is limited by weak intu...

The article considers a hybrid approach to anomaly detection taking into account the parameters of network endpoints. The transformation of the network endpoint parameters to the format of the session parameters is implemented using regrouping and dimensionality reduction. To detect anomalies, ensembles of classifiers are used, with the inputs rece...

This article discusses the recognition and solution of social dilemmas in multi-agent reinforcement learning. A multi-agent environment has been developed that simulates the movement of cars on a circular two-lane road, with two agent cars independently trained using a neural network. For effective training, the algorithm of independent Q-learning...

In recent years, reinforcement learning methods have become increasingly important for many applied areas. Such learning assumes the presence of a reward function. The reward should be the higher, the more the agent’s behavior corresponds to the desired one. At the same time, in many cases, the reward function is built axiomatically, by expert sele...

Optimization of mobile computing is a relevant and promising direction, since there is currently a request to change the operating system-application paradigm, as well as a request to reduce the amount of information in popular mobile ecosystems. The article considers a new area of manifestation of the fundamental laws of the evolution of adaptatio...

The bulk move of employees to a remote mode of work, as well as the intensified information confrontation, led to the emergence and aggravation of many problems related to security in corporate networks. Cases of attacks on local area networks, in particular, using phishing and social engineering techniques, distribution of malicious code, and DDoS...

The extremely large number of possible states in the real environment limited the implementation of reinforcement learning algorithms until recent times. Hence, such algorithms were implemented only for simple tasks in tabular grid worlds. Deep learning have expanded the use of such algorithms for multidimensional and complex virtual environments o...

To avoid information systems malfunction, their integrity disruption, availability violation as well as data confidentiality, it is necessary to detect anomalies in information system operation as quickly as possible. The anomalies are usually caused by malicious activity – information systems attacks. However, the current approaches to detect anom...

The article discusses the ways to overcome the difficulties associated with surfing the Internet for certain information. Many users are not able to explicitly make up a search request according to their information needs, and their search is often limited to haphazard use of several most popular search engines. Most search engine research is techn...

Many real-world problems contain social dilemmas that express contradictions between individual and public interests. Such tasks can be solved using deep multi-agent learning. This article focuses on reputation-based social dilemma solving and multi-agent reinforcement learning. Target of the work is the development of a method for solving social d...

Redesigning of the whole portal or some of its webpages does not always lead to a positive result despite the designers' expectations. Nowadays the opinion on how to carry out the redesign of web portals varies. A web portal being in service and in a wide use, a sudden change in its design may result in a drop of the number of clients. The usabilit...

Today, there is no consensus about proper timing and conditions for integration of PowerPoint presentations into the educational process. But the model-based evaluation can make a decision-making process easier when it comes to using presentations. The purpose of this study is to build a formal model to evaluate presentations. In order to build a f...

Network technologies have been steadily developing and their application has been expanding. One of the aspects of the development is a modification of the current network attacks and the appearance of new ones. The anomalies that can be detected in network traffic conform with such attacks. Development of new and improvement of the current approac...

Nature-inspired algorithms have come into use to solve more and more optimization tasks of high dimension when classical optimization algorithms do not apply. The task of user interface usability optimization becomes the one to be solved by nature-inspired algorithms. Usability optimization suggests a choice of interface design out of a large numbe...

Recently, human recognition systems based on deep machine learning, in particular, on the basis of deep neural networks, have become widespread. In this regard, research has become relevant in the field of protection against recognition by such systems. In this article a method of designing a specially selected type of camouflage applied to clothin...

The rapid growth of computing devices has led to the emergence of distributed user interfaces. A user interface is called distributed if a user can interact with it using several devices at the same time. Formal methods for designing such interfaces, in particular methods for the distribution of interface elements across multiple devices, are yet t...

In the last few years, human recognition systems based on machine learning, specifically on deep neural networks, are on the rise. This, in turn, increased researchers' interest in protection against such systems. In this paper, we propose a method of designing camouflage patterns that will protect a person from being recognized by a human observer...

This paper concerns moving object tracking in the videos, based on sparse optical flow technique. Current optical flow tracking methods suffer from feature points loss. We extended an existing sparse optical flow tracking method with a new function for automatic feature points recovery that uses biological regeneration principle. Besides, we improv...

This article describes how the conversion rate of a web page depends on the interface usability degree. Optimization of existing interfaces as the matter of improving their usability faces a number of difficulties. In the first place, the unified objective function selection method for such optimization is not set up; that is resulting in necessity...

Human re-identification is a crucial component of security and surveillance systems, smart environments and robots. In this paper a novel selective covariance-based method for human re-identification in video streams from multiple cameras is proposed. Our method, which includes human localization and human classification stages, is called selective...

An automatic quantitative method for speech fragment analysis and chronological ordering is proposed. The method works by first converting the audio data into two-dimensional spectrograms and then extracting a large set of 1030 numerical descriptors (features) from the raw spectrograms, as well as from transforming the spectrograms. The audio fragm...

Extended model of goals, operators, methods and selection rules (GOMS) for quantifying gesture interfaces suggested. User experience of interacting with gesture interface, controlled by hand motions, analyzed. The components and coefficients of extended model are proposed, experimentally tested and verified. The mathematical description of extended...

The data fusion technology allows one to combine heterogeneous information coming from different sources in digital application areas. Criteria aggregation operators are one of the effective ways to implement data fusion. But selecting the appropriate aggregation operator for a specific application is a great challenge. Nevertheless, there are meta...

The Choquet integral based on fuzzy measure is a generalization of the weighted average operator. Due to flexibility in the aggregation of information and ability to take into account possible uncertainties in data the integral has been gaining popularity in the application of multicriteria decision making. In this paper an application of fuzzy int...

This paper discusses the methodology of usability testing on the base of the real-time analysis of user’s electroencephalogram (EEG). In the experimental study of the control groups of users working with a real interface of commercially relevant applications, two distinct user states were revealed which correspond to the superposition of EEG beta a...

Information retrieval based on weighted zone scoring means the assignment weight for each zone or each field in the document metadata. All these weights are obtained using machine learning methods. The paper presents a method of determining the weights using the fuzzy Choquet integral. This allows taking into possible account interdependence betwee...

By analyzing muscle activity with the BioPlux device multiple user gestures can be recognized. Gestures are independently selected from each other. Relevance of the topic from the practical point of view is determined by the need to create a software system that can use sign language interface in real time. The main objective of this study is to cr...

A logic programming approach to the intelligent monitoring of anomalous human activity is considered.The main idea of this approach is in using of a first order logic for describing abstract concepts of anomalous human activity, i.e. brawls, sudden attack, armed attack, leaving object, loitering, pickpocketing, personal theft, immobile person, etc....

A primary goal of virtual environments is to support natural, efficient, powerful and flexible human-computer interaction. But the traditional two-dimensional, keyboard- and mouse-oriented graphical user interface is not well-suited for virtual environments. The most popular approaches for capture, tracking and recognition of different modalities s...

Development of systems for intellectual interaction in general and intelligent multimodal interfaces in particular involves employment of specific methodologies for development of an specific architecture composed of a converter of perceptions and situations, modifiers of models of user’s behaviour and the environment, recognizer of situations, as...

Choquet integral with respect to fuzzy measure is a generalization of weighted arithmetic mean aggregation operator. It allows taking into account the phenomenon of dependence between criteria. Due to this it is possible to reflect the expert knowledge more accurately without making the model simplification which is the assumption of independence o...

A method of data fusion in multimodal interface using fuzzy fusion operators is described in this paper. Special scenes are determined by dynamic and static patterns for which the sources of information are the results of pattern recognition by low-level algorithms

Intelligent multimodal interface is user interface that can provide users with the opportunity to interact adequately to their modalities, recognize different situations to achieve users’ goals, create a model of user’s behavior and the environment in which it operates, modify behavior in accordance with the current state of the user and the enviro...

Intelligent human-machine interfaces based on multimodal interaction are developed separately in different application areas. No unified opinion exists about the issue of what properties should these interfaces have to provide an intuitive and natural interaction. Having carried out an analytical survey of the papers that deal with intelligent inte...

This article addresses Web personalization based on the analysis of individual user activity. However, human behavior is characterized by uncertainties that should be considered in the personalization algorithms. Fuzzy logic allows taking into account different types of uncertainty. Therefore, the paper presents a method for Web personalization bas...

Most of the data today is unstructured: books, newspaper articles, publications, reports, and especially images, audio and video streams. The people need the special interface to deal with the huge amount of unstructured data. Especially it is needed to develop approaches to recognize the semantic of incoming video information. The paper proposes s...

The paper includes approach on the base of multimodal interface with the environment through the use of hierarchy model of fuzzy finite state machines (FSM). FSMs used to define the relationships between objects extracted from video streams in the process of an environment observation and to define the different dynamic situations.

Having carried out an analytical survey of the papers that deal with intelligent interfaces the set of properties are presented, which are necessary for intelligent interface between an information system and a human

A primary goal of virtual environments is to support natural, efficient, powerful and flexible human-computer interaction. But the traditional two-dimensional, keyboard- and mouse-oriented graphical user interface is not well-suited for virtual environments. The most popular approaches for capture, tracking and recognition of different modalities s...

Nowadays wide-spread opinion that instead of training human users to be computer-literate in order to use the technology, the technology itself should be made intelligent enough to communicate and understand users in a human-friendly way. Therefore speech and gesture interfaces (including the recognition of facial expression, hand and body gestures...

A method of optimal fuzzy aggregation of secondary attributes in problems of recognizing a person from images is proposed. The primary attribute is a characteristic feature of an object, allowing to distinguish one object from another. The secondary attribute is a result of recognition algorithm, based on primary attributes. Fuzzy integrals of Suge...

In this paper, teleoperation system in which a robot performs a block of commands based on gestures commands sent through a wireless network, is described. An original method of manipulative gesture recognition is proposed. The method is based on presentation of gestures as fuzzy Finite-State Machines and its equivalence grammars. Recognition consi...