ABSTRACT. The problem of modeling the emotions in an artificial cognitive system is considered within the Natural-Constructive Approach worked out in our previous works (BICA 2013, 2015, 2016). Especial attention is paid to interpretation and imitation of aesthetic emotions caused by impression of Art, Music, Natural Phenomena (rainbow, sunset, etc.). In contrary to so called pragmatic emotions (associated with certain pragmatic goal), the aesthetic emotions have no rational reasons and appear to be quite individual and often inexplicable, i.e., could not be strictly formulated and explained. We consider these problems using the Natural-Constructed Cognitive Architecture (NCCA), which represents the complex multi-level hierarchical composition of different-type neural processors (in continual representation based on the original dynamical formal neuron concept). The whole system is divided into two separated (but linked) subsystems in analogy with two cerebral hemispheres. It is shown that emotions, being induced by certain sub-cortical structures (thalamus, amygdale, etc.), provide the tool that should control the ‘dialog’ (mutual interaction) between the subsystems in course of solving certain problems. The aesthetic emotions are shown to be connected with some indirect and unformulated associations provided by the lowest (image) level of hierarchy (fuzzy set). It is shown that the concept of chef-d’oeuvre and corresponding effect of goose bumps could be caused by the paradox of recognition arising when the object seems familiar and unusual simultaneously.

ABSTRACT. A wide family of machine learning algorithms are often called adaptive or data-driven methods, due to their capability of adaptation to the available set of data, learning by example, requiring no physically grounded analytical or computational model or a priori knowledge of the studied object. Such methods may be often called biologically inspired – as by their origin, as by resemblance of their behaviour to the behaviour of data processing systems in living creatures. The scope of problems that are solved by such methods includes those of prediction, evaluation, classification, clusterization, inverse problems, and other data analysis problems. Examples of such methods are artificial neural networks (ANN) and the method of partial least squares, or projection to latent structures (PLS). From mathematical point of view, these methods are sophisticated approximation methods using adaptively tuned combinations of relatively simple functions of most general type. Many physical methods are based on indirect measurements, and therefore they imply solution of inverse problems (IP) – determination of the sought-for parameters by the observed values. Such problems are often ill-conditioned or even incorrect. That is why adaptive methods of IP solving based on approximation of the inverse function are demanded and efficiently used. Attention is driven to the main differences between ANN and PLS, and the main shortcoming of PLS – that it is a linear method (yet it is the best linear method). Even with an adequate non-linear pre-processing of data, PLS is often unable to build an approximation comparable by its quality with that implemented by an ANN. The main advantage of PLS is its low computational cost. In the lecture, methodological aspects of using ANN are discussed. From the point of view of data processing methods, any IP can have various formulations: as a regression, classification (for discrete-valued IP) or optimization problem. The key differences of ANN as a method of solving IP from alternative methods are discussed. When solving IP, ANN can be used within one of several methodological approaches: “model-based”, “experiment-based”, and “quasi-model”. The difference among these approaches, their properties and areas of application are described. A separate question arises if the IP being solved is a multi-parameter one. The possible approaches to the order of determination of parameters are autonomous determination, simultaneous determination of all parameters, group determination (with joining of parameters into groups with simultaneous determination within each group), and stepwise determination (when some of the parameters already determined are used as additional inputs for determination of other parameters). Other useful additional methods discussed in the lecture are cluster-based approach, when the problem domain is separated into several sub-domains, and the IP is solved separately in each of these sub-domains, and training with adding noise into training data, thus increasing noise resilience of the solution. It is stressed that with increasing complexity of a problem, linear methods begin to fail, and ANN turn out to be more resilient to this increasing complexity. The general purpose of the lecture is to attract attention of a wide audience of young scientists to the great opportunities opened by use of biologically inspired adaptive methods, and by the latest methodological achievements in IP solution by ANN. The material is illustrated by examples of IP from two areas of physics – optical spectroscopy and electrical prospecting.

ABSTRACT. The trend of recent years has been the advent of 2D and 3D cryptographic transformations. Stand-ards that have appeared in the 21st century, specify algorithms based on the use of 2D and 3D transformations (AES, Kuznechik, Keccak, Stribog). In the article a 3D version of cryptographic transformation specified by GOST 28147-89 is suggested. The 3D GOST algorithm is characterized by the high degree of parallelism at the level of elementary operations. Increasing bit depth of the processed data blocks from 64 to 512 bits al-lows 3D GOST to be used for the synthesis of hash algorithms. Algorithm improvement agenda may be similar to the DOZEN family of algorithms.

ABSTRACT. This article describes a communication system that protects user data against theft or damage. In addition to any standard communications system elements the described communication system includes a security module. The security module is installed on the data transmission bus between the processor and communications. The security module includes a processing unit, capable of handling data according to a particular algorithm: encryption, masking or other. It eliminates the possibility of the transmission of any data to communication modules due to undocumented features of the processor, or any other module of the mobile device. The operating system and the security module only know the processing algorithm. Consequently, no information, other than that sent to the operating system leaves the device. The security module provides security and confidentiality of data; it does not require the production of a special security processor or any other units and can be implemented on the element base of leading world manufacturers.

ABSTRACT. In this talk I will provide a focused overview of the representational assumptions developed in different cognitive architectures. In doing so I will outline the main problematic aspects of the current proposals (involving the limited size and the homogeneous typology of the encoded and processed knowledge) and I present a possible way out based on different, but complementary, research agendas.

ABSTRACT. Beginning with the works of gestalt psychologists (Köhler, 1921; Wertheimer, 1959) the existence and the role of insight – the key moment in problem solving, associated with an abrupt reorganization of a problem representation, which leads to its solution and is often accompanied by vivid emotional experiences – was questioned. For all its recognizability (for example, most solvers confidently report that they experienced it when solving certain problems), the existence of insight is by no means universally recognized. Substantiated doubts in the reality of this phenomenon appeared almost immediately after the quoted work of V. Kohler. Thus, already in 1949, D. Hebb described acute long-standing disputes between those who were confident in the existence of insight (configurationists), and those who denied the necessity of attracting this concept (learning theorists) (Hebb, 1949). The situation changed drastically after the introduction of the problem space theory by A. Newell and H. Simon (Newell, Simon, 1972) who proposed a step-by-step approach to the goal state with associated gradual local changes of a problem’s representation instead of a single-step solution discovery. This report is intended to show the ambiguity of the concept of insight and to characterize the present state of its theoretical and experimental research.

ABSTRACT. WHAT SORT OF CONSCIOUSNESS MACHINES CAN HAVE? This is an application of non-reductive materialism to AI. I will present new material how to recognize machine consciousness (the epistemic problem) based on the work on analysis of sex robots in my paper "Church-Turing Lovers" (OUP, upcoming in an anthology by Lin et al., 2017) and on quantum effects (based on recent work by Sky Darmos and others); creativity aspect extended based on recent research by Stephen Thaler and Troy Kelley. Early project presented in: (2012) "The Engineering Thesis in Machine Consciousness" Techne: Research in Philosophy and Technology 16(2), 187-207 and (2009) "The Philosophical Issue in Machine Consciousness," International Journal of Machine Consciousness 1(1), 155-176 PDFs available at: https://sites.google.com/site/peterboltuc/publications

Invited by Alexei Samsonovich

ABSTRACT. This paper proposes a method of automatic VLSI circuit analysis. On the first step transistors are grouped by their structure. Groups with irregular structure are highly interconnected to each other. Detecting Tangled Logic Structures (TLS) with a GTL-depended linear ordering and genetic algorithm divides the circuit due to its functional structure and forms the gate-level VLSI circuit. High-level functional blocks in circuit description consist of gate-level cells groups, which are also highly interconnected. After TLS-blocks extracting, it is possible to describe their function. TLS-blocks are smaller, represent a cell of high-level circuit, and are thus more suitable for further functional circuit analysis than a gate-level VLSI circuit.

ABSTRACT. Modern views on visual attention mechanisms, some results of our last psychophysical experiments with eye movements recording and future steps of studying have been considered. Several groups of finding have been determined, namely: (i) unresolved objectives; (ii) problems for which contradictory data are known; (iii) finding which propose revision of some classic views; (iv) views consistent with data obtained in different research centers. The results of psychophysical tests directed on receiving of quantitative parameters of eye movements at performance of different visual tasks to estimate the contribution of bottom-up and top-down mechanisms of visual attention are presented. It is supposed that obtained experimental results can be formalized to use in realistic mathematical models of visual attention.

ABSTRACT. Learning machine understand natural human speech is significantly easier and more effective when understood the cognitive processes in humans. Psycholinguistics, as an interdisciplinary study of language and cognition, is extremely helpful in this respect. The aim of this study is to understand what linguistic coping strategies people develop under stress due to their physiological and psychological structure. An oral communication in a foreign language, English, of people with different backgrounds is being investigated under an isolation stressor. It is expected that the physiological and psychological features, adrenaline and noradrenaline value indications and anxiety tendency, influence the linguistic performance in a stressful situation. Subjects who share similar organization of the central nervous system and psychological structure, may also indicate similarities in cognition regarding the language production. The value of the study for artificial intelligence research is in its applicability on natural language systems, e.g. by helping program a more precise speech recognizing system.

ABSTRACT. The main goal of this paper is to describe new integrated methodology: Agent-based Reconfigurable Generic Organization. This is an integrated approach which allows designing future product life cycle organizations as a class of learning, self learning or adaptive distributed systems using high-level, sign-based communications across entire product lifecycle environment named Product Related Semiosphere (PRS). The definition of PRS is a fundamental question of developing a new class of dstributed semiotic systems with ability to communicate, identify and manufacture engineering artifacts with prescriptive characteristics. The paper analyses different types of product life cycle management approaches and suggests overall approach for integration of business and engineering knowledge during the whole product life-cycles. It allows to understand the interrelations of different life-cycle stages for acquiring and manipulating concurrent engineering knowledge. The authors proposed the idea of using socio-inspired framework based on applied semiotics and distributed artificial intelligence.

ABSTRACT. A body of knowledge shows the emerging of an evidence according to a better account for the emotion spectrum is achievable by employing a detailed selection of emotion keywords. Basic emotions, such as Ekman’s ones, cannot be considered universal, but are related to with implicit thematic affairs within the corpus under analysis. The paper tracks some preliminary experiments obtained employing a data-driven methodology that captures emotions, relying on domain data that you want to model. The experimentation consists of investigating the corresponding conceptual space based on a set of terms (i.e., keywords) that are representative of the domain and the determination. Furthermore, the conceptual space is exploited as a bridge between the textual content and its sub-symbolic mapping as an “emotional fingerprint” into a six dimensional hyperspace.

ABSTRACT. This paper presents the conceptual framework for a study of musical experience and the associated architecture centred on Human-Humanoid Interaction (HHI). We discuss the state of the art of the theoretical and the experimental research into the cognitive capacity of music. We overview the results that points to the correspondence between the perceptual structures, the cognitive organization of sounds in music, the motor and affective behaviour. On such grounds we bring in the concepts of musical tensions and functional connections as the constructs that account for such correspondence in music experience. Finally we describe the architecture as a models generator system whose modules can be employed to test this correspondence from which the perceptual, cognitive, affective and motor constituents of musical capacity may emerge.

ABSTRACT. Most artificial general intelligence (AGI) system developers have been focused upon intelligence (the ability to achieve goals, perform tasks or solve problems) rather than motivation (*why* the system does what it does). As a result, most AGIs have an unhuman-like, and arguably dangerous, top-down hierarchical goal structure as the sole driver of their choices and actions. On the other hand, the independent core observer model (ICOM) was specifically designed to have a human-like “emotional” motivational system. We report here on the most recent versions of and experiments upon our latest ICOM-based systems. We have moved from a partial implementation of the abstruse and overly complex Wilcox model of emotions to a more complete implementation of the simpler Plutchik model. We have seen responses that, at first glance, were surprising and seem-ingly illogical – but which mirror human responses and which make total sense when considered more fully in the context of surviving in the real world. For ex-ample, in “isolation studies”, we find that any input, even pain, is preferred over having no input at all. We believe that the fact that the system generates such un-expected but “humanlike” behavior to be a very good sign that we are successful-ly capturing the essence of the only known operational motivational system.

ABSTRACT. The goal of this work is to study the influence of the neuron model choice on the results of STDP learning on base of simple toy tasks. As shown, the resulting mean output firing rate after STDP learning with restricted symmetric spike pairing scheme does not depend on the mean input rates for such neuron models as Leaky Integrate-and-Fire, Traub, and static neuron. Then this effect, being used to solve a typical classification task of Fisher’s Iris, demonstrates that the classification accuracy does not depend significantly on the choice of the neuron model. Thus, the independence of learning results on the neuron model gives the possibility to use simpler neuron models in further investigations.

ABSTRACT. We present a model of Reinforcement Learning, which consists of modified neural-network architecture with spatio-temporal connections, known as Temporal Hebbian Self-Organizing Map (THSOM). A number of experiments were conducted to test the model on the maze solving problem. The algorithm demonstrates sustainable learning, building a near to optimal routes. This work describes an agents behavior in the mazes of different complexity and also influence of models parameters at the length of formed paths.

ABSTRACT. The construction of biologically inspired cognitive architectures is based on data obtained by studying the mechanisms of the brain's functional networks operations, the causality of their integration and differentiation into the neurophysiological architectures of the cognitive processes of consciousness. One of the key networks involved in maintaining the basic level of consciousness at the resting state is the default mode network (DMN). The network activation increases with perceived mental states as imagination, internal dialogue, and others and reduces with an external stimulation or behavioral task. A complete loss of consciousness is characterized by the synchronization of DMN with the anticorrelated with DMN network. When the level of consciousness changes in the processes of cognitive activity, there is a complex picture of the combination of positive and negative connections between different networks and regions of the brain. At the same time, the changes in the intrinsic brain organization during the cognitive process and the resting state still open question. This work was aimed at studying the dynamics of the architecture of different brain networks interactions at the resting state, including executive and attention networks, cerebellum, DMN, visual network, auditory network, brainstem, the somatosensory and motor networks, sub cortical network. Three algorithms were used for clustering states in neural network connectivity dynamics: direct clustering of functional network using k-means algorithm, modularity-based clustering, topological based clustering . The obtained results showed that in the dynamics of functional neural network connectors there are three expressed states, determined by different types of interactions between DMN networks, attention and other neural networks.

ABSTRACT. Grid cells found in the mammalian medial entorhinal cortex (mEC) are neurons with spatially modulated receptive fields that tile 2D space in a hexagonal pattern. Together with place cells, head direction cells and boundary vector cells, grid cells are thought to constitute a neural system for the encoding and navigation of space. The spacing between successive grid fields increases in discrete steps along the dorso-ventral axis of mEC. This property, along with the periodic nature of their firing patterns, suggests that grid cells may provide a high capacity encoding of space. Moreover, grid cells are thought to update their firing patterns primarily by path integration (PI), the process of integrating self-motion in order to maintain a continuous estimate of location. However, navigation by path integration alone is impractical since noise is integrated over time, leading to drift away from the correct location. Accordingly, the stability of grid firing patterns are known to degrade slowly in darkness, but remain in the presence of sensory cues, suggesting that these serve to stabilise location estimates arising from path integration. However, the grid firing pattern can also deviate from its otherwise uniform hexagonal arrangement, becoming more elliptical in large square environments, deforming in trapezoidal environments, and adopting a less periodic arrangement on a 1D track. Grid firing patterns also rescale parametrically in response to the contraction or expansion of a familiar environment and reconfigure into globally consistent firing patterns when the animal is allowed to navigate between two previously unconnected environments. Existing models of grid cells use sensory input to reset the grid firing pattern, rather than probabilistically integrating this input with the PI estimate. Here, we propose that probabilistic integration combined with plasticity between the grid cells and their sensory inputs can account for many of the deformations described above. We also show that integration of these two cues can produce stable grid patterns where both signals are uncertain when considered independently. These results highlight the role of sensory inputs in shaping our perception if space and challenge the assumption that grid cells provide a universal metric of space.

ABSTRACT. Trust pervades nearly every social aspect of our daily lives; it penetrates all human interactions from personal relationships to organizational interactions. A plethora of studies have started to gain a deeper understanding of the inherent nature of trust by combining behavioral paradigms with functional neuroimaging, electroencephalographic, lesion, endocrinological, and genetic methods in both healthy and psychopathological populations. However, an overarching framework that integrates those separate findings from different levels into a conceptual framework characterizing trust behavior is still lacking. In this summer school, I will sketch out an integrative neuroscience framework on the neurobiological underpinnings of trust that provides an explanation on how trust behavior emerges from the interplay of genes, hormones/ neurotransmitter, brain circuits, and cognition. The integration into a unified conceptual framework will guide future investigations of the complex interplay between both biological and environmental factors and facilitate the understanding of humun-human but also human-machine inerpersonal trust.

ABSTRACT. The paper considers the problem of visualizing heterogeneous information relevant to the solution of a particular problem domain. An essential part of the task is to get the conversion of data objects doing their representation adjusted for the corresponding data model. Creation of a model of converting of data objects is offered on the basis of applicative computing systems. Achievement of flexibility requires the parametrization of the considered construction, i.e. support of dependence of a set of available methods of interpretation on parameters as which semantic characteristics of processed data appear. The methods of working with interpretation coordination tools have been partially tested when implementing various applications for informational support for the implementation of the best available technologies (BAT).

ABSTRACT. As the Internet of Things evolves, the attack surface - the number of vulnerable points - will become greater. The huge number of things connected to the Internet will create opportunities for DDoS attacks, identity theft, money theft, etc. This means that we are on the threshold of a new era, which will usher in new cyberthreats and new challenges for security systems developers. The emergence of sophisticated complexes of things that interact with each other and the need for context-based and behavioral analysis mean that cognitive technologies will play a crucial role in ensuring Internet of Things security. Neuromorphic computing systems are also of great interest: given the right architectural solutions, they can contribute to the development of a world of things that will be immune to attacks thanks to their bioinspired intelligence.