next day
all days

View: session overviewtalk overview

10:00-11:30 Session Poster1: All posters will be on display in all poster sessions. Posters listed below should be presented by authors in this session only.
On the Regularity of the Bias of Throughput Estimates on Traffic Averaging

ABSTRACT. The interaction of intelligent agents implies the existence of an environment to support it. The usual representations of this environment are graphs with certain properties. Like reliability, throughput is one of the most important characteristics of such graphs. When evaluating throughput in the analysis and synthesis of graphs, a reasonable combination of heuristic and strict approaches is used. In practice, this leads to the use of graph metrics. Usual shortest paths are widely used as part of the various multi-colour flow distribution procedures. The analytical capabilities of the Euclidian metric can achieve much more than just obtaining such a distribution. Such a metric allows us to introduce an abstract measure of the (quadratic) proximity of an arbitrary graph to a complete graph. This measure can be used as a single indicator of the reliability and throughput of the graph. Other conditions for tasks on graphs can be attributed to restrictions. A traffic matrix is one of these conditions. Non-stationarity of traffic when averaging over time can significantly reduce the accuracy of the estimates of throughput. The described dependencies of the throughput on the traffic's non-stationarity can be used in the analysis and synthesis of the communication environment when organizing the structure of the interaction of intelligent agents in the conditions of limited resources. These dependencies are verified by the results of numerical experiments.

Lightweight Surround View algorithm for embedded TDA3xx platform
PRESENTER: Margarita Zaeva

ABSTRACT. Automotive surround view camera system is an emerging automotive ADAS (Advanced Driver Assistance System) technology that assists the driver in parking the vehicle safely by allowing him/her to see a top-down view of the 360 degree surroundings of the vehicle. In our work, we propose a lightweight solution for camera automatic calibration with preconfigured distortion parameters and single-shot optical center and external parameters configuration. In addition, we propose synthesis algorithm to work with new calibration optimized for DSP c66x with 25fps.

Development of an AI recommender system to recommend concerts based on microservice architecture using collaborative and content-based filtering methods
PRESENTER: Andrey Malynov

ABSTRACT. Recommender system is a complex software system primarily intended to select the most relevant content based on user's personal preferences. In order to achieve the set goal, a number of tasks must be completed, including: • track user actions across various devices; • get product data from a number of sources and maintaining their currency; • consolidate the data; • create user and product profiles bases on big data in a real-time mode; • select recommendations in cold start and highly sparsed data environment; • assess the quality of the recommender system. Completion of each specific task must not extend the time to complete other tasks. Users must instantly get the relevant content even if the system is heavily loaded, for example, due to a popular event announcement. A workaround may be to divide the system into independent components with the ability to scale specific services. Microservice architecture examined in this article intends to ensure required flexibility due to asynchronous message exchange via a data bus and other principles offered by SOA concept. Apart from interaction between the components, the article also introduces the results of development of each specific service from asynchronous user action tracker to recommender engine based on the hybrid approach that includes collaborative and content-based filtering methods, and the knowledge-based approach using Artificial Intelligence techniques. Special attention is paid to a subject category with a number of aspects that prevent applying generic approaches to building recommender systems.

Preliminary Experiment on Emotion Detection in Illustrations Using Convolutional Neural Network

ABSTRACT. The paper describes an experiment on emotion detection in images, specifically illustrations and cartoon images. Usually, detection and classification of emotions are performed on human faces so the algorithm can learn, for example, what a “happy human face” looks like. These algorithms probably can’t transfer their understanding of happiness features onto different types of objects, like animals or cartoon illustrations. We, humans, can recognize and detect signs of emotions (although maybe falsely) in new and unusual objects. Developing an algorithm capable of recognizing emotions in objects it wasn’t trained on would allow for better human-like robots and systems. This is a preliminary study on how well knowledge gained by a typical neural network detection system on a set of objects transfers to new, unknown objects. The neural network detection system used in this study is YOLO. We collected small training datasets using cartoon illustrations of several animals of two categories: happy and sad. We tested the trained network on a set of illustrations depicting a different animal the network hasn’t seen in training. The best performance achieved is 0.69 F1-score.

Creating a model using blockchain for swarm of biobots guidance
PRESENTER: Igor Prokhorov

ABSTRACT. This paper is about a relatively new concept will be proposed, in which blockchain technology will be used to operate swarm of biobots and store information about their movement. In previous articles, we talked about how you can create an automatic biobot control system using deep neural networks based on the slitherin artificial intelligence framework. Next, we will consider a model for applying the combined approach of blockchain technology and the automatic control system of biobots. The application of this concept can open up new opportunities for optimization in swarm robotics.

Cognitive model of the project's environment in case of Russian mega-projects abroad

ABSTRACT. Russian megaprojects abroad are complex and costly programs for the creation of large industrial and energy facilities. One of the most important tasks that arise in the framework of such a project is information support at all its stages, which includes tracking the semantic field of media and social media publications and managing emerging information risks. Modern risk management assumes the presence of a mathematical and statistical model that allows you to predict the dynamics of processes in the subject area served. This paper presents new models of the megaproject environment based on the use of fuzzy cognitive maps (FCM). A special feature of these models is the hierarchical principle of organizing FCM concepts. The resulting maps are used to predict a number of business metrics. To build fuzzy maps, we use the developed original tools. This work was supported by RFFI grant № 20-010-00708​\20

Instrumental and algorithmic tools for constructing and verifying fuzzy cognitive maps based on the example of maps that model nuclear industry enterprises

ABSTRACT. Creating new predictive models of behavior of complex socio-economic systems based on machine learning methods is one of the most relevant areas of development in economic and mathematical modeling. A promising approach to solving this problem is the use of neuro-fuzzy systems, which can be obtained by combining fuzzy cognitive maps and neural network models. This work is devoted to the development of tools for fuzzy cognitive modeling. The methods developed by the authors for automatic construction, training, and verification of cognitive maps are described, including those based on neural network training methods. The models of the external and internal environment of the nuclear industry enterprise built in the software application developed by the authors are presented.

A method for reducing the impact of information risks on a megaproject life cycle based on a semantic information field

ABSTRACT. This paper consider a comprehensive method for devising loyalty programs based on the stages of a life cycle of an international megaproject. The method is based on the analysis of information risks and their management. The method is focused on aggregation and processing of data from various sources of textual information, which demonstrates the attitudes of key categories of individuals regarding the implementation of a megaproject at its numerous life cycle stages. The semantic information field is formed using hardware and software based on Neural Network Technologies. Authors examine the most popular neural network architectures that are used in the sentiment analysis. The paper describes a comparative analysis of classification accuracy of neural network architectures based on volume of texts and their ability to sentiment analysis of large and small volumes of text. The method is aimed at managing and influencing information flow that accompanies the implementation of a megaproject stages. The application of semantic information field makes it possible to account for the informational risks of a megaproject and to prepare an effective set of measures to counteract these risks in a timely manner. This work was supported by RFFI grant № 20-010-00708\20.

Algorithmic and instrumental tools for thematic modeling and annotation of texts based on convolutional neural networks

ABSTRACT. Automated processing and analysis of text data, along with the rapid growth of their volume, is becoming more and more relevant for the development of machine learning methods. Among these tasks, it is worth highlighting such as: topic modeling aimed at extracting topics from the corpus of texts, as well as automatic annotation and extraction of keywords to describe these texts. This paper presents algorithms and their software implementation aimed at solving these problems, based on the use of modern convolutional networks, such as BERT, cluster analysis methods and graph algorithms. The results of experimentation on various data, including scientific articles and texts of any subject, are described.

Prediction of time series values based on a neuro-fuzzy system

ABSTRACT. Time series of values of economic metrics are an important source of information when making decisions. The development of predictive models for such series is an extremely urgent task. Despite the existence of numerous methods for solving this problem, a universal approach has not yet been developed. Moreover, given the rapidly changing factors that affect the dynamics of the economic situation, both at the global and local levels, the development of accurate and adequate forecast models is in demand in any industry. This paper presents original algorithms for predicting some macro-economic indicators by combining fuzzy cognitive maps and neural networks, both direct propagation and recurrent, into a single model. The results of experiments both on artificially generated data and on open data extracted from social media are described.

Complex Objects Identification and Analysis Mechanisms
PRESENTER: Mikhail Ulizko

ABSTRACT. Currently the volume of information on the Internet global network is inclined to increase. It affects various areas of activity. The paper presents the issue of identifying and analyzing complex objects in the context of information sources devoted to project activity of the US National Institutes of Health (NIH). As part of the solution to the problem, information sources relevant to the provided topic were found, and the data were downloaded with the use of agent techniques. The methods of primary analysis and data processing were developed to create a data storage structures in SQL and NoSQL models. The analysis of the presented database models was conducted, their advantages and disadvantages were revealed. As a result software tools have been developed that provide data representation of a complex object and organization of work with it by web interface.

Cyber Polygon Site Project in the Framework of the MEPhI Network Security Intelligence Center

ABSTRACT. At present, the market for information protection tools (IPTs) is much wider than a couple of years ago. But not only technology protects and carries a threat. People are still at the forefront as the most common cause of errors is the lack of experience and low competency. The only right solution is the creation of cyber polygons as specially equipped and controlled network infrastructures for developing practical skills to combat information security (IS) threats. The National Research Nuclear University MEPhI (Moscow Engineering Physics Institute) could not remain aloof from this process as the leading institute for IS training in Russia. Therefore, it was decided to create such a cyber polygon within the framework of the educational and research Network Security Intelligence Center (NSIC) for intelligent network security management es-tablished at the MEPhI Institute of Cyber Intelligence Systems in 2016. The pa-per describes the first results achieved in making this project a reality. It intro-duces the “Cyber Polygon” term, briefly analyzes a state of the current cyber polygons development worldwide, and introduces the MEPhI Cyber Polygon objectives and provision to be used within the framework of the “Business Con-tinuity and Information Security Maintenance” Master’s degree programme. Further activities in its development conclude the paper.

Network Security Intelligence Centres for Information Security Incident Management

ABSTRACT. Intensive IT development is driving current information security (IS) trends and require sophisticated structures and adequate approached to manage IS for different businesses. The wide range of threats is constantly growing in modern intranets; they have become not only numerous and diverse but also more disruptive. In such circumstances, organizations realize that IS incidents’ timely detection and prevention in the future (what is more important) are not only possible but imperative. Any delay leaves only reactive actions to IS incidents, putting assets at risk as a result. A properly designed IS incident management system (ISIMS), operating as an integral part of the whole organization’s governance system, reduces IS incidents’ number and limits damage caused by them. To maximally automate IS incident management (ISIM) within one organization and to deepen its knowledge of IS level, this research proposes to unite together all advantages of a Security Intelligence Centre (SIC) and a Network Operations Centre (NOC) with their unique and joint toolkits and techniques in a unified Network SIC (NSIC). This paper presents the research, which is focused upon the designing and evaluating the concept of NSICs, and represents a novel advancement beyond existing concepts of security and net-work operations centres in current security monitoring scenarios. Key contributions are made in relation to underlying taxonomies of threats and attacks, leading to the requirements for NSICs, the related design, and then evaluation in a practical context and the implications arising from this (e.g. training requirements).

Visualization of T. Saati hierarchy analysis method
PRESENTER: Dmitriy Kovtun

ABSTRACT. The article is devoted to the problem of expert assessment visualization of the comparative importance of various indicators. T. Saati hierarchy analysis method, which allows us to obtain indicators' weight coefficients by paired comparisons of them, is considered. A method for increasing the accuracy of estimates formed by an expert is proposed. The increase in accuracy is achieved by instantly visualizing the results of the comparisons for their prompt adjustment by an expert. This work was supported by RFFI grant № 20-010-00708\20.

“Loyalty Program” tool application in megaprojects
PRESENTER: Anna Tikhomirova

ABSTRACT. The article is devoted to the problem of improving the implementation efficiency and megaprojects management. An example of an adaptation one of the existing project management methodological tools, namely the “loyalty program”, for projects of a global nature (megaprojects) is given. The authors propose the use of elements of the decision-making process theory and the method of T. Saati hierarchy analysis method for the analysis of megaproject key characteristics. Based on them, recommendations for the formation of a loyalty program are given. This work was supported by RFFI grant № 20-010-00708\20.

Block Formation for Storing Data on Information Security Incidents for Digital Investigations
PRESENTER: Andrey Nikiforov

ABSTRACT. Nowadays technologies such as Blockchain (BC) and the Internet of Things (IoT) can be heard everywhere. But because of the leap in the development of these technologies, there is a need to evaluate the existing approaches critically. One of the up-to-date tasks is to study information security (IS) incidents as a part of the IoT. Due to a large number of different manufacturers and options for technology implementation, it cannot be unambiguously concluded what choice will be better. The paper examines the related work in the area and proposes an approach to form the basis for storing data on IS incidents. For this purpose, the authors formulate a block structure for including in a chain for later use, for example, in computer forensics.

Method of Applying Fuzzy Situational Network to Assess the Risk of the Industrial Equipment Failure
PRESENTER: Margarita Zaeva

ABSTRACT. The method of formation and application of a fuzzy situational network for the solution of problems of an assessment of risk of failure of the production equipment of the industrial enterprise at statement on production of a new product is offered. The method takesintoaccountthemaingroupsofmanagementactionsinsmallandmedium-scaleproductionofawiderangeofproducts.The condition of the equipment is represented in the form of a fuzzy situational network with a three-level structure. The following levels are distinguished: situational, network and factor.

Anthropogenic Spatial Systems: Transformation and Potential of Self-Organization
PRESENTER: Margarita Zaeva

ABSTRACT. The paper analyzes and summarizes the results of the study of the features of the functioning and transformation of a particular type of spatial systems, namely anthropogenic. These systems are formed for different purposes of human activity. Refinements of basic definitions are offered and it is shown that stability of functioning for such systems is the most significant indicator of their existence. The ability of anthropogenic systems to self-development and the influence of self-development on their transformation under the influence of internal and external influences and factors are considered. Proposals for the application of methodological foundations of system diagnostics of the state and transformation of anthropogenic spatial systems, as well as the possibility of their algorithmization and information support are presented.

The method of planning the process of refueling vehicles using artificial intelligence and fuzzy logic methods
PRESENTER: Margarita Zaeva

ABSTRACT. Filling liquefied natural gas (LNG) into the vehicle’s onboard tanks is one of the most dangerous production operations when using LNG as a motor fuel. Technical requirements form the volume of hazardous substance (LNG) stored at the refueling facility and restrictions on fire distances (gaps) and measures, which leads to an increase in the facility’s area and the cost of its construction and operation. One of the ways to reduce costs associated with protective measures is to limit the amount of fuel stored at the facility, while providing the flow of vehicles with fuel with its limited reserves at refueling facilities. To solve this problem, it is proposed to use the planning of the process of refueling vehicles using fuzzy logic methods using artificial intelligence. As a result, the fueling company’s fuel information service provides drivers with recommendations on possible refueling locations, fuel costs, available fuel volumes, and waiting times.

Linear systems theoretic approach tointerpretation of spatial and temporal weights incompact CNNs: Monte-Carlo study
PRESENTER: Artur Petrosyan

ABSTRACT. Interpretation of the neural networks architectures for decoding the signals of the brain usually reduced to the analysis of temporaland spatial weights. We describe a theoretically justified method of their interpretation within the novel architecture based on a priori knowledge of the subject area. This architecture is comparable in decoding quality to the winner of the BCI IV competition and allows to for automatic engineering of physiologically meaningful features. To demonstrate the operation of the algorithm, we performed Monte Carlo simulations andreceived a significant improvement in the restoration of patterns for different noise levels and also investigated the relation between the decoding quality and patterns reconstruction fidelity.

11:30-12:00 Session Plenary1: Mostly Eastern Hemisphere
An Adaptive Temporal-Causal Network Model to Analyse Extinction of Communication over Time

ABSTRACT. The persistence of information communicated between humans is difficult to measure as it is affected by many features. This paper presents an approach to computationally model the cognitive processes of information sharing to describe persistence or extinction of communication in Twitter over time. The adaptive mental network model explains, for example, how an individual can experience information overflow on a topic, and how this affects the sharing of information. Parameter tuning by Simulated Annealing is used to identify characteristics of the network model that fit to empirical data from Twitter. The data collected is related to the independentism in Catalunya, Spain, which is considered a global issue with repercussion in Europe.

Extending affective capabilities for medical assistive robots
PRESENTER: Giovanni Pilato

ABSTRACT. In this work, we discuss methodologies and implementation choices to enable a humanoid robot to estimate patients’ moods and emotions during postoperative home rehabilitation in the context of the AMICO project.

14:00-16:00 Session Plenary2: For All Participants
Modeling the Emergence of Informational Content by Adaptive Networks for Temporal Factorisation and Criterial Causation

ABSTRACT. Propagated activation of neurons through their network is often considered as the main process in the brain. However, another crucial part of neural processing concerns adaptation of characteristics of this network such as connection strengths or excitability thresh-olds. This adaptation can be slow, as in learning from multiple experiences, or it can be fast, as in memory formation. These adaptive network characteristics can be considered informational criteria for the activation of a neuron. This then is viewed as a form of emergent information formation. Activation of neurons is determined by such information, termed criterial causation. In the current paper, the relationship of criterial causation with the principle of temporal factorisation for the dynamics of the world in general, describing how the world represents information about its past in its present state, which then, in turn, determines the world’s future. In the paper, it is shown how these processes were analysed in more detail and modeled by (adaptive) network models.

Modeling the Development of Internal Mental Models by an Adaptive Network Model
PRESENTER: Raj Bhalwankar

ABSTRACT. Mental models and the mental process of modeling them play a crucial role in individual and organizational learning. The adaptive mental processes involved in the development of mental models are addressed here. The reported study integrates a number of psychological and neurological theories on mental models and the learning involved. An adaptive network model has been designed for these processes and used for simulations addressing a case study of learning to drive a car. The developed model may be valuable for different ends, like in improving individual and organizational learning, in designing virtual pedagogical agents, enhancing driver safety, and self-driving systems in Cars.

Multi-modal Actuation with the Activation Bit Vector Machine

ABSTRACT. Research towards a new approach to the abstract symbol grounding problem showed that through model counting there is a correspondence between logical/linguistic and coordinate representation in the visuospatial domain. The logical/verbal description of a spatial layout directly gives rise to a coordinate representation that can be drawn, with the drawing reflecting what is described. The main characteristic of this logical property is that it does not need any semantic information or ontology apart from a separation into symbols/words referring to relations and symbols/words referring to objects. Moreover, the complete mechanism can be implemented efficiently on a brain inspired cognitive architecture, the Activation Bit Vector Machine (ABVM), an architecture that belongs to the Vector Symbolic Architectures. However, the natural language fragment captured previously was restricted to simple predication sentences, with the corresponding logical fragment being atomic Context Logic (CLA), and the only actuation modality leveraged was visualization. This article extends the approach on all three aspects: adding a third category of action verbs we move to a fragment of first-order Context Logic (CL1), with modalities requiring a temporal dimension, such as film and music, becoming available. The article presents an ABVM generating sequences of images from texts.

Learning algorithm for an intelligent decision making system based on multi-agent neurocognitive architectures
PRESENTER: Zalimhan Nagoev

ABSTRACT. The paper presents the formalism of an intelligent decision-making system based on multi-agent neurocognitive architectures, which has an architectural similarity to the human brain. An invariant of the organizational and func-tional structure of the intellectual decision-making process based on the mul-ti-agent neurocognitive architecture is developed. An algorithm for teaching intelligent decision-making systems based on the self-organization of the in-variant of multi-agent neurocognitive architectures is presented. Using this algorithm, an intelligent agent was trained and the architecture of the learning process was built on the basis of an invariant of neurocognitive architecture. Further research is related to training an intelligent agent in more complex behavior and expanding the capabilities of an intelligent decision-making system based on multi-genic neurocognitive architectures.

Simple Model of Origin of Feeling of Causality

ABSTRACT. A simple computer model of a feeling of causality of autonomous agents has been created and investigated in the current article. The model of the evolution of a population of agents is considered. The population includes agents of two kinds: 1) agents with a feeling of causality and 2) agents without such a feeling. Each agent has its internal resource. Agents with a feeling of causality remember causal relationships between situations in the external environment. It is shown that agents with a feeling of causality in the process of evolution can displace agents without a feeling of causality from the population. So, the current model can be considered as the model of origin of a feeling of causality.

Conceptual Processing System for a Companion Robot

ABSTRACT. Companion robots should perceive speech, recognize objects in the real world, and further react with speech utterances and nonverbal communicative cues. Robots should also remember the interaction history and accumulate knowledge from different text sources: news, blogs, and e-mails. We de-sign a conceptual representation system for a companion robot, able to sup-port this list of interactive tasks. The system includes speech processing component and operates with semantic representations as sets of semantic markers, assigned to valencies. The reaction support system inherits a classic production architecture and consists of scripts, sensitive to rational or emotional stimuli. The general architecture is based on parallel processing of scripts, it may trigger several behavioral reactions to each stimulus, and further combine the outputs of these scripts on a robot to enrich its communicative behavior. Semantic representations and scripts are also used to index in-coming utterances are store them in a memory base. We also demonstrate interaction of semantics and reactions with a prototype of visual recognition system for the tasks of face detection and automatic support of Tangram puzzle solution.

BICA Society Panel

ABSTRACT. Every year at the BICA conference, a BICA Society Panel is held as a special session dedicated to the society meeting. The BICA Society Board of Directors uses this opportunity to report on its activities and achievements during the past year and to discuss new decisions and plans for the future, to summarize statistics of the current event and discuss plans for the next BICA conference. To talk about current and future publication channels. to discuss any business related to the BICA Society Membership (including Member benefits) and activities, including online repositories. To vote, when necessary, for new Directors and/or to ratify any amendments to the Bylaws. And, the last but not the least, to issue awards. This year's agenda may be abridged due to the online-only format of the conference; at the same time, the panel may be distributed into several short sessions. BICA Society is a nonprofit 501(c)(3) organization, incorporated in the state of Delaware in 2010, with its headquarters currently located in Seattle, WA. The officers of BICA Society are the three Directors who are listed as the authors of this abstract. BICA Society is also an international scientific community founded in 2010, that currently includes hundreds of Members (including inactive), is associated with and supported by the BICA Society corporation. The mission of the BICA Society is to promote and facilitate the transdisciplinary study of Biologically Inspired Cognitive Architectures (BICA), in particular, aiming at the emergence of a unifying, generally accepted framework for the design, characterization, development, implementation and evolution of human-level cognitive architectures. The specific goal put forward by BICA Society in 2010 is known as the BICA Challenge. It is the challenge to implement computationally the most essential higher human cognitive abilities, such as the Self and agency, free will, social emotions, the ability to understand what's happening, the ability to learn actively like a human, etc., making them work in real-life conditions, thereby creating an example of a Strong Artificial Intelligence. Therefore, we value all contributions from all fields and research areas using one criterion: how much do they advance us toward a solution of the BICA Challenge.

16:30-19:30 Session Plenary3: For All Participants. This session will also include other speakers and the BICA Society Panel.
Recent Research on the Soar Cognitive Architecture and Interactive Task Learning

ABSTRACT. General autonomous intelligent agents with ongoing existence have many challenges when it comes to learning. On the one hand, they must continually react to their environment, focusing their computational resources and using their available knowledge to make the best decision for the current situation. On the other hand, they need to learn everything they can from their experience, building up their knowledge so that they are prepared to make decisions in the future. We posit two distinct levels of learning in general autonomous intelligent agents. Level 1 (L1) are architectural learning mechanisms that are innate, automatic, effortless, and outside of the agent’s control. Level 2 (L2) are knowledge-based learning strategies that are controlled by the agent's knowledge, whose purpose is to create experiences for L1 mechanisms to learn from.

We describe these levels and provide examples from our research in interactive task learning (ITL). In ITL, an agent learns novel tasks through natural interactions with an instructor.ITL is challenging because it requires a tight integration of many of the cognitive capabilities embodied in human-level intelligence: multiple types of reasoning, problem solving, and learning; multiple forms of knowledge representations; natural language interaction; dialog management; and interaction with an external environment – all in real time. Our agent is implemented in Soar, and uses a combination of innate L1 mechanisms and L2 strategies to learn ~60 puzzles and games, as well as tasks for a mobile robot. Our agent is embodied in a table top robot, a small mobile robot, a Fetch robot, and Cozmo.

A Multi-Level Cognitive Architecture for Self-Referencing, Self-Awareness and Self-Interpretation

ABSTRACT. In this paper, a multilevel cognitive architecture is introduced that can be used to model mental processes in clients of psychotherapeutic sessions. The architecture does not only cover base level mental processes but also mental processes involving self-referencing, self-awareness and self-interpretation. To this end, the cognitive architecture was designed according to four levels, where (part of) the structure of each level is represented by an explicit self-model of it at the next-higher level of the architecture. At that next-higher level, states reify part of the structure of the level below; these states have a referencing relation to it. In this way the overall architecture includes its own overall self-model. The cognitive architecture was evaluated for a case study of a realistic type of therapeutic session from clinical practice.

Toward Integrating Cognitive Components with Computational Models of Emotion using Software Design Patterns
PRESENTER: Enrique Osuna

ABSTRACT. Computational models of emotion (CMEs) are software systems designed to imitate particular aspects of human emotions. The main purpose of this type of computational model is to capture the complexity of the human emotion process in a software system that is incorporated into a cognitive agent architecture. However, creating a CME that closely imitates the actual functioning of emotions demands to address some challenges associated with the design of CMEs and cognitive architectures alike. Among these challenges are i) sharing information among potentially independently developed cognitive and affective components, and ii) interconnecting complex cognitive and affective components that must interact with one another in order to generate realistic emotions, which may even affect agents' decision making. This paper proposes an architectural pattern aimed at cataloging and describing fundamental components of CMEs and their interrelationships with cognitive components. In this architectural pattern, external cognitive components and internal affective components of CMEs do not interact directly but are extended by including message exchange methods in order to use a publish-subscribe channel, which enables their intercommunication, thus attenuating issues such as software heterogeneity. This structural approach centralizes communication management and separates the inherent complexity of the cognitive and affective processes from the complexity of their interaction mechanisms. In so doing, it enables the design of CMEs' architectures composed of independently developed affective and cognitive components. The proposed architectural pattern attempts to make progress in capturing the complex process of human emotions in a software system that adheres to software engineering best practices and that incorporates quality attributes such as flexibility and interoperability.

The architecture of voluntary action in biological and synthetic brains

ABSTRACT. Volitional motor control can be seen as the result of a gradual replacement of feedback by feedforward control. We have addressed this question from the perspective of an integrated architecture called the Distributed Adaptive Control (DAC) theory of mind and brain. DAC proposes that the brain is a multi-layer control system which optimizes the how of action by considering why (motivation), what (objects), where (space), when (time) and who (agents) or the H5W problem. We have shown that for DAC to realize optimal solutions in foraging problems, its decision-making renders policies that simultaneously optimize perceptual evidence, memory bias, goals, and utility. This raises the question of what the principles are that underlie the processing and adaptation of these factors. In this presentation, I will focus on a link between policy adaptation and perceptual learning we have recently advanced. The dominant model of anticipatory motor control relies on the notion of an inverse model that by learning from encountered errors acquires corrective responses that supersede feedback control. However, these models are predicated on a Markovian world assumption and thus by necessity face problems in handling exceptions, such as observed in probe trials, where fast feedback control is required. We solve this challenge by proposing that adaptive goal-oriented motor control can also be obtained by relying on a cascade of purely sensory predictions that drive feedback control via counterfactual errors or Hierarchical Sensory Predictive Control. At the highest level this includes the predictions derived from the goal-oriented decision-making by the embodied agent. This proposal provides a new interpretation of the neuronal systems underlying volition and agency and motor learning. Using robot experiments, we have demonstrated the robustness of this solution. We have found further supporting evidence for the relevance of counterfactual error in the physiology of motor learning, the neurophysiology of human memory and in the rehabilitation of stroke patients. In addition, with direct electrophysiological recordings from intracranially implanted pharmacoresistant epilepsy patients we have found direct evidence for the virtualization underlying the notion of counterfactual error.

References: Ballester, B. R., … (2016). Counteracting learned non-use in chronic stroke patients with reinforcement-induced movement therapy. Journal of neuroengineering and rehabilitation. Ballester, B., … (2019). A critical time window for recovery extends beyond one-year post-stroke. J. Neurophysiology. Maffei, G., … (2017). The perceptual shaping of anticipatory actions. Proc. R. Soc. B. Pacheco, D. P., Sánchez-Fibla, M., Duff, A., Principe, A., Rocamora, R., Zhang, H., ... & Verschure, P. F. (2019). Coordinated representational reinstatement in the human hippocampus and lateral temporal cortex during episodic memory retrieval. Nature communications, 10(1), 1-13. Verschure, P. F., … (2003). A real‐world rational agent: unifying old and new AI. Cognitive science. Verschure, P. F., Voegtlin, T., & Douglas, R. J. (2003). Environmentally mediated synergy between perception and behaviour in mobile robots. Nature.

Robot passes the mirror test by inner speech
PRESENTER: Antonio Chella

ABSTRACT. To make a robot able to pass the mirror test is a well-known research problem. The existing strategies are based on kinaesthetic-visual matching and require to manipulate perceptual data. The proposed work attempts to demonstrate that is possible to perform robust self-recognition on the basis of self-dialogue which manipulates just verbal information. By labelling signals, it is possible to conceptual reason on them and to solve the problem of self-recognition. The idea is supported by the existing literature in psychology, where the importance of inner speech in self-reflection and in self-concept emergence was empirically demonstrated.

Virtual Convention Center: A socially emotional online/VR conference platform

ABSTRACT. A cognitive model, producing believable socially emotional behavior, is used to control the Virtual Actor behavior in an online scientific conference paradigm. For this purpose, a videogame-like platform is developed, in which Virtual Actors are embedded as poster presenters, receptionists, etc. The expectation is that the combination of somatic factors, moral appraisals and rational values in one model has the potential to make behavior of a virtual actor more believable, humanlike and socially acceptable. Implications concern future intelligent cobots and virtual assistants, particularly, in online conferencing and distance learning platforms and in intelligent tutoring systems.

20:00-21:00 Session Poster2: All posters will be on display in all poster sessions. Posters listed below should be presented by authors in this session only.
Cognitive Modeling and Analysis for Multiple Interacting Users in Collaborative Knowledge-Building
PRESENTER: Anamika Chhabra

ABSTRACT. People neither behave uniformly in their social lives nor is their behavior entirely arbitrary. Rather, their behavior depends on various factors including their skills, motives, and background. Our work shows that such a behavior also prevails in the websites of Stack Exchange. We collect and analyze the data of over 5.3 million users from 156 Stack Exchange websites between the year 2008 and 2016. In these websites, users' diverse behavior shows up in the form of different activities that they choose to perform as well as how they stimulate each other for more contribution. Using the insights gained from the empirical analysis as well as the classical cognitive theories, we build a general cognitive model depicting the users' interaction behavior emerging in collaborative knowledge-building setups. The analysis of the model indicates that for any given collaborative system, there is an optimal distribution of users across its activities that leads to the maximum knowledge generation. We also apply the model on Stack Exchange websites and identify the under-represented activities.

To Help or Not to Help: A Network Modelling Approach to the Bystander Effect

ABSTRACT. This paper focuses on defining and simulating behavioural outcomes of the bystander effect. These insights were modeled by temporal-causal networks. Typical patterns of bystander behaviour were translated into three requirements and seven simulated scenarios of the by-stander effect. All scenarios were simulated to showcase the main bystander effect dynamics and its accordance with the literature. Unknown parameters of the effect were further esti-mated by a Simulated Annealing algorithm. In the end, the created model shows the potential to simulate the bystander effect in different and new scenarios.

Study of neurocognitive mechanisms in the concealed information paradigm
PRESENTER: Sergey Kartashov

ABSTRACT. This work is a continuation of research aimed at creating a forensic method for MRI diagnostics of hidden information in a person. The article presents some results of research in the paradigm of information concealment, during which the activity of brain structures in experiments with tests used in criminology was studied using fMRI and MRI-compatible polygraph (Mrtsp). In numerous experiments it is shown and confirmed the effective-ness of MRI in the course of the fMRI and the usefulness of the methodo-logical technique created through the use RTSP and improve the perfor-mance of an fMRI experiment.

Development of a laboratory workshops management module as part of a Learning Support System for the "Decision-Making Theory" course

ABSTRACT. The article is devoted to the problem of optimizing the load of teachers in the educational process. The task of developing a specialized Learning Support System with respect to specifics of the subject is considered. The authors propose a universal mechanism for managing laboratory workshops as a tool for monitoring and evaluating the degree of "Decision-Making Theory" course material assimilation.

Comparison of brain induced potentials in internal speech in studied and unknown languages
PRESENTER: Alisa Suyuncheva

ABSTRACT. Abstract. Significant differences were found in the bioelectric responses of the brain to words with the same meaning in Russian and Japanese in different se-ries in the electrophysiological experiment on a sample of 18 people. Two groups were used: those who knew Japanese and Russian, and those who knew only Russian as their native language. Differences in mutual articulation of these words were found due to differences in evoked potentials of the late latency (300-500 MS) in the frontal, central and parietal leads, and also in the earlier - 200 MS - latencies in the central, temporal and frontal leads). When comparing between groups (knows Japanese or does not know), differences in the components of EP 200 and P300 were obtained by speaking in Russian, howev-er, there is no difference in primary and semantic processing (P100, N400). . The main significant differences over the channels С3, F3, P3 given in this chapter are the amplitude and latency of latency in N200, as well as the de-crease in amplitude N400. These data are correlated with earlier studies.

Reflection of Empathy Processes in Evoked Potentials

ABSTRACT. Empathy is the ability to recognize, understand, and share the feelings of another person, including the perception of another person. Empathy pro-motes prosocial or helping behavior that comes from within, rather than under duress. Our paper revealed brain activity including empathy and altruistic processes in evoked potentials (EP). To review empathy, we analyze the brain activity characteristics with electrodes during watching the empathic game “Stone-Paper-Scissors”. Results based on the nineteen-channel electroencephalography (EEG) re-cordings experiment on a sample of 16 participants (6 women and 10 men). Using the EP method, electrical activity was measured for the situation when the empathy behavior was activated. The evoked potential consists of a sequence of negative and positive deviations from the mainline and lasts 500ms after the end of the stimulus. In EP, evaluate the amplitude and la-tent period of occurrence. To register the EP, the same electrodes are used as for the EEG recording, and with the unified observational conditions. We focused on brain structures associated with prosocial behavior, including the cortex, amygdala, and thalamus. A comparison between the em-pathy situations and the ‘first-person’ experience has been performed separately for reactions of men and women. We found significant differences in the following leads: T6, P4, O2, Pz in EP obtained in 2 series separately by gender.

The electroencephalogram based classification of internally pronounced phonemes

ABSTRACT. The internal speech recognition is a promising technology, which could find its use in brain-computer interfaces development and greatly help those who suffer from neurodegenerative diseases. The research in this area is in its early stages and is associated with practical value, which makes it relevant. It is known that internal pronunciation can be restored according to electroencephalogram data because it allows one to register specific activity associated with this process. The purpose of this work is to build and implement an algorithm for extracting features and classifying Russian phonemes according to an electroencephalogram recorded during the internal pronunciation of the phonemes. This kind of research is actively conducted abroad; however, at the moment, there is no information in open sources about such works for the Russian language phonemes. In the course of the work, an algorithm for extracting features and classifying the internal pronunciation of Russian phonemes was built and tested, the accuracy of which showed results comparable with other studies.

The Role of Gender in the Prosocial Behavior Mechanisms

ABSTRACT. Prosocial behavior is progressively being studied because of its enormous role in the changing conditions of the modern world. The behavior influences distant communication conducts Artificial Intelligent and causes psychological disorders, such as autism spectrum disorder or psychopathy. We study the prosocial behavior by analyzing the brain features mechanisms during the custom setting - empathy game “Stone-Paper-Scissors”. Observed results for the cohort of 55 participants (28 women and 27 men) were obtained from the test questionnaires and EEG. During the experiment 29 distinct brain structures were analyzed by the author’s “virtual implanted electrode” method. In the method, electrical activity was measured for the situation when the empathy behavior was activated. Our approach relates to the field of medicine and neuroscience, in particular, to a method for studying the activity of individual brain structures predefined by their spatial position according to the scalp multichannel EEG. A comparison between the empathy situations and the ‘first-person’ experience has been performed separately for gender reactions of men and women. The results of the questionnaires were tabulated and processed by the factor analysis. Six factors were identified that accounted for 41.45% of the total variance of the data. The factors received following interpretation: “Altruistic”, “Publicity”, “Emergency prosocial behavior”, “Conformist prosocial behavior”, “Anonymous prosocial behavior”, “Emotional prosocial behavior”. Comparison by Student's T-test showed a significant (p <0.05) difference between men and women in the “Publicity” factor (men are more public than women) and in the “Emotional prosocial behavior” factor (females are more emotional than males).

A Methodology for Designing Self-organizing Systems

ABSTRACT. In existing multi-agent systems, agents are used that perform various, but strictly regulated processes, and their inducing occurs according to a schedule specified by the operator or according to some preset conditions. Such approach leads to a high level of downtime of computing power in solving problems of preparing information and analytical data in a short time, be-cause agents are not universal enough, and their incentive to work is not event-oriented. The article presents a methodology for group multi-agent problem solving. Group autonomous problem solving is possible only if there is autonomous group management and, in particular, the possibility of self-organization, i.e. changing of structural units within a group and changing of its behavior depending on the actual state of the external environment. The practical solution to self-organization problems in agent systems seems to be the most important direction for increasing their intellectual characteristics and the areas of their practical use.

Changing educational content based on analysis of user behavior in distance learning systems
PRESENTER: Sergey Nemeshaev

ABSTRACT. Distance learning systems have become a convenient and useful tool in the educational process. Teachers can quickly place training materials, prepare test questions, create tests. Using such systems allows you to get detailed statistics on the assimilation of material by students, identifying the topics that caused the most problems in passing the tests. However, analysis of only the final results gives only a general idea of the student's success. In this article we will show that analysis of additional metrics, such as response time for each task, changing the solution when choosing an answer, skipping questions and then returning to them, and other metrics, can improve the quality of analysis of material assimilation and identify problem areas in training.

Using virtual reality technologies in the educational process
PRESENTER: Sergey Nemeshaev

ABSTRACT. Recently, new teaching methods are gaining popularity. About 10 years ago, the distance learning platforms were actively launched - it gave the opportunity to learn at a convenient time, in a convenient place and at the same time to choose the best courses and receive knowledge from recognized experts in their science area. Today we see that with the advent of new technologies, new opportunities open up - learning with virtual or augmented reality. In our article we will consider not only the main advantages of learning process with the use of these technologies, but also propose methods that will improve educational content based on the analysis of students' behavior.

Post-quantum group key agreement scheme
PRESENTER: Julia Bobrysheva

ABSTRACT. Progress in quantum technologies forces the development of new cryptographic primitives that are resistant to attacks of an adversary with a quantum computer. A large number of key establishment schemes have been proposed for two participants, but the area of group post-quantum key establishment schemes has not been studied a lot. Not so long ago, an isogeny-based key agreement scheme was proposed for three participants, based on a gradual increase in the degree of the key. We propose another principle for establishing a key for a group of participants using a tree-structure. The proposed key establishment scheme for four participants uses isogeny of elliptic curves as a mathematical tool.

Multiagent model of the process of mastering linguistic competence based on perceptual space formation
PRESENTER: Irina Gurtueva

ABSTRACT. We propose a simulation model of the early development of language competence. In its core it is a model of phonemic imprinting that describes the process of mapping audio stimuli into classes of elementary units of a language, taking into account the influence of social factors. The machine learning algorithm used in the model was developed after the results of a study of the features of speech used in referring to children. This model will allow us to explore the features of phonetic perception, the cognitive mechanisms that underlie language development, highlight the main factors affecting the duration of the period of plasticity. On that basis, we can build perceptual maps, create diagnostic tools to describe the sensitive period, which will facilitate the study of the stages of its opening and closing. The model can also be used to create speech recognition systems that are resistant to a variety of accents and effective when used in a noisy environment.

Eye movement correlates of foreign language proficiency in Russian learners of English

ABSTRACT. Today, the modeling of various aspects of speech activity is the mainstream of modern cognitive and computational science. Along with models of natural language processing, much attention is paid to finding mechanisms of the several languages functioning within one linguistic system (bilingual, trilingual, etc.). The search for specific features of the processing of linguistic information by one subject in different languages allows one to approach the construction of bilingual systems models. This article is devoted to the analysis of the eye movements in reading texts in native and foreign languages by bilinguals with different levels of proficiency in the latter. The present study tests the assumption that eye movement features of people with a high level of foreign language skills are similar during text reading in native and foreign languages. Another goal is to elicit features that provide the differentiation between the elementary and the intermediate levels of English language proficiency. We offer new Eye Tracking based evaluation metrics for the level of language proficiency.