Lecture
Introduction
Our laboratory is studying the behavioral mechanisms of the human psyche, the principles of the formation of the motivational hierarchy and its functional apparatus. The goal of our work is to build a self-governing system capable of not only responding to external stimuli like a living being, but also adapting a complex of its needs and goals to its own capabilities and habitat. To accomplish this task, we needed to develop and integrate an element of intentionality into the control system (for details of the definition, see “Theoretical Base”). Without it, any model would be no more than a trivial repetition of the well-known implementations of the stimulus-response algorithm (with an initially specified set of needs-attitudes of the simulated organism). From here, the algorithms of the “solution search” implemented in the system are only instrumental and auxiliary ones. The leading role is assigned to the block of formation of individual goals (attitudes) - the block that controls the evolution of needs and forms its hierarchy (motivational structure) in the modeled subject.
During the implementation of the task, a technology of motivation was developed, on the basis of which the tool software was created. These products can be in demand:
1. Researchers and developers in the field of Artificial Intelligence;
2. Psychologists studying behavior and adaptation processes, behaviorists, etc .;
3. Developers of an Artificial Personality with elements of intentionality in behavior (academic studies and applied developments, for example, in the entertainment and robotics industry);
4. Sociologists and HR-specialists, (monitoring interpersonal conflicts, etc.).
For each of the listed aspects of the application, the essence of the technology “Melos” and its capabilities will be described below.
The theoretical basis of modeling was formed under the influence of the works of famous mathematicians and philosophers, who gave the definition of Artificial Intelligence (AI) and developed a whole trend in cybernetics and modeling of Artificial Personality. First of all, this is Alan Turing, Franz Brentano, Daniel Dennett, and many others. Some of their theses reflect key provisions of the concept adopted as the basis for the construction of the Melos system. For example, it is worth mentioning the interpretation of the essence of intentionality (in particular, the unfolding function of goal formation) as a distinctive feature of mentality (F. Brentano). In the works of D. Dennett, who gave more substantive formulations of the provisions of Brentano, signs of intelligence in the behavior of the system (organic or mechanical origin) were the following:
a) the possibility of interpreting behavior in the context of the environment in which it takes place;
b) the integrity of all elements of behavioral reactions and their integration into a general system subordinate to one or another goal (the criteria of holism);
c) the consistency of the sequence of all behavioral patterns of the system with the general task of the most optimal adaptation.
Further, Dennett states: “If an artificially created mechanism reveals behavior that is subordinate to the general task of its existence, it is expedient in its aspect, it is consistent and is in accordance with the changing conditions for the implementation of this task - this behavior can be characterized as possessing intentionality, having rationality as far as rationality is manifested in our own activities. " Intentionality, according to Dennett, entirely and exclusively depends on a certain type of system's behavioral patterns, and it does not matter how organized and in what form their internal mechanisms are organized.
This approach was adopted by our laboratory when creating the mechanism of motivation (as controlling the entire behavioral apparatus of the model). However, the wording of philosophers and cybernetics only outlined the main goals. The detailed development of the elements of the motivational mechanism and the construction of the mathematical apparatus were carried out on the basis of the work of a number of psychologists and sociologists. To create a model of the motivational mechanism, formalization of theoretical positions and practical observations by R. Burns, A. Maslow, C. Rogers, D. Uznadze, Z. Freud, E. Fromm and many others was required. Their work formed the basis of a mathematical model that allows setting on one hand the characteristics of the social environment, on the other, the innate abilities of the individual, after which, in the process of modeling, observe the dynamics of changes in the individual's motivational sphere that occur during the evolution of a given subject in a modeled environment.
The principles of building a logical apparatus of the model were ideas about the motivational system as a hierarchical structure of energy nodes, subject to certain sewage laws of the stimulus energy and controlling (and also controlled) its generation and distribution.
The approach to the analysis of the psyche from the standpoint of its informational structure was suggested by the works of L. M. Wecker, consideration of the mechanism of motivation from the perspective of the hierarchical system was made in accordance with A. Maslow’s textbook postulates, and the sewage system of motivational energy is based on ideas expressed in the works of G. Murray .2
1 Quotes from Abstracts A. Turing, F. Brentano, D. Dennett are cited in the book “Mind Design II” (Philosophy Psychology Artificial Intelligence (revised and enlarged edition)) © 1997 Massachusetts Institute of Technology; A reference list of the work of psychologists is contained in the full description of the technology "Melos".
2 Under the motivational apparatus in this material is meant a complex of mental mechanisms, the cumulative activity of which can be considered in the aspect of a unified system that controls the motivation in the human psyche (stimulus - goal - act). The concept of “environment” unites all the factors that influence this system from the outside and are independent of the modeled subject. Naturally, other (contacting with the subject) individuals whose behavior influences his motivational attitudes are also included in the environment. These individuals can be subjects of parallel modeling, but they can also act as extras for the “social background” task.
The main functional core in the modeling system is a unique technology of self-regulation of motivational knots, realizing the mechanism of achieving optimal pleasure and avoiding frustration. The activity of this mechanism is governed by the Ego-concept apparatus, which independently develops the hierarchical structure of an individual's needs in accordance with his abilities, capabilities, and conditions available for the activities of this individual environment (set of contexts). The combination of these mechanisms opens up a wide spectrum for using both the concept of modeling itself and software created on its basis in a number of fields (both academic and applied). Below is a list of the main uses of technology.
Solutions and algorithms developed during the creation of the “Melos” system are implemented in the form of business logic blocks. These blocks can be integrated into the control system of the behavior of any mechanism (both real robots that mimic the behavior of animals or humans, and electronic characters "living" in a virtual environment). This will allow to give their behavior qualitatively new elements, including:
a) the mechanism of self-learning of the subject in accordance with his personal needs and the surrounding conditions for their implementation;
b) a unique multi-factor system for finding the optimal solution to the problem (satisfying needs with regard to the changing conditions of their implementation context), as close as possible to the algorithms of the activity of a living person;
c) the mechanism for choosing the most preferable motivations (voluntary goal setting), taking into account the personal experience of the subject, his individual abilities and habitat (context), as well as other attributes of intellectual activity.
In addition, it becomes possible to introduce purely psychological qualities and characteristics in the description of the AI model: accentuations, aptitudes, abilities, attitudes - what gives the subject's activity a unique individuality. At the same time, each individual trait evolves and transforms both under the influence of other installations and under the influence of “habitat”. Thus, the individual experience of the adaptive activity of the subject forms his system of motivation (direct feedback). This regulation of the subject's behavior is subordinated to the single task of his self-realization, including the independent choice of instrumental goals and methods of their achievement, as well as being integrated into the habitat context in which it takes place.
Created on the basis of the Melos technology, instrumental software allows the development of behavioral models based on randomly selected and modified hierarchies of the needs of the modeled subjects. The internal structure of the developed software is not tightly tied to any of the well-known systems of descriptions of needs (Maslow, Murray, etc.). Instead, the user is provided with a set of developed templates, which he can change at his discretion or apply them in the original form. The user can also develop their own systems of motivational hierarchy. The software provides the main thing - the interface for manipulating objects of the motivational structure and all the internal mechanisms of their interactions. The following list of functionalities gives an idea of the capabilities of the Melos simulation system in this aspect of its application:
a) a formalized representation of the elements of the motivational structure, their detailed parametrization, a description of their behavior and changes over time, including correlations with each other, as well as self-organization of the interaction of motivational elements under the control of the parameters included in their description;
b) the construction of multilevel (decomposable) motivational structures, for example, the abstract motivational setting “conformity to the cultural requirements of society” is capable of decomposing into substantive actions objectifying the degree of this conformity;
c) the organization of a flexible (experimentable configurable) hierarchy of needs suitable for most theories of motivation (the presence of ready-made templates to facilitate the first steps);
d) an assessment of the priority of the planned actions for making efforts to implement the most important motive in the current conditions, as well as working off unrealizable motives in the imagination (psychological defense function);
e) tracing of all transactions of motivational objects - both internal (sewage motifs within the subject) and external (integration of the motivational object accepted from the outside into the subject's hierarchy and subsequent reorganization of channels within it), as well as maintaining the state of the motivational structure in a static form for subsequent analysis or management process.
It should be added here that only the most important functions of the system are listed here in this aspect of its application. Thanks to the unified technical base on which the tool software is implemented, the system fully utilizes all its capabilities (including those described in other paragraphs).
The implementation of an Artificial Personality directly echoes the task of building Artificial Intelligence, but in this case the emphasis falls not on the rational side of mentality, but on its emotional qualities. In this case, the elements and attributes of the human behavior of an artificial system are not so much the ability to self-study and improve their activities, as the subject’s behavioral patterns depend on his personal experience, initial properties and accentuations, on the subjectivity of his attitude to the surrounding motivators (presented in the context of modeling) and other similar factors. The modules implemented in the Melos system that form and support this subjectivity make it possible to use it both in academic studies (allowing integration of these modules with other systems) and in application programs using human behavior imitation algorithms (entertainment industry and robotics). Among the most significant features are the following:
a) description of the habitat, including an unlimited number of subjects whose motivational systems experience mutual influence in the process of evolution;
b) dynamic and autonomous formation of individual tendencies in the modeled subject, their development into full-fledged motivational knots;
c) the mechanism of training and socialization of the subject, which, on the one hand, is completely controlled by the experimenter (shaping the initial conditions of development), on the other hand, is revealed completely autonomously, as a fragment of the general adaptation apparatus of the psyche.
Finally, the psychosocial aspect of using the Melos system should be mentioned. Tools created on the basis of its technology, allow to model not only the behavior of individuals, but entire groups. Moreover, the most adequate model is the one in which behavior correlates with social factors. It is these correlations that may be of interest when analyzing small (relatively closed) social groups in the aspect of complementarity of the motivational vectors of their constituent individuals. The functionality of the system includes the following algorithms:
a) self-determination of the authority of the persons surrounding the subject (included in the modeled context of evolution), affecting their assimilation of certain stimuli or patterns of behavior;
b) a dynamic change in the self-esteem of the subject, based both on the results of their own claims / achievements, and on comparisons with other subjects of the environment (context).
All this makes it possible to realize a large set of sociological experiments that experience the reactions of a typical subject in varying situations (social environment, general living conditions, etc.).
Using the appropriate template, using the instrumental software, it is possible to calculate the optimality of the organization of motivating vectors in relatively closed communities and groups (formation of working groups, intra-family correlations, etc.), their distribution among the subjects involved in a single process (activity pursuing common goals). A new direction is opening up in pedagogical experiments with a maximum approximation to real individuals, for example, the removal and transfer of a child’s motivational portrait to a formalized representation of a modeling system. The list of applications of the system in this aspect can be extended by analyzing the effects on the inner world of the subject of means of external influence (arbitrary advertising, violent incentives, habitat canons, etc.). It also becomes possible to determine the motivational integrity and completeness of the context of the subject’s habitat - in terms of its needs and the possibilities for satisfying them (in other words, determining the adequacy of the habitat to its personnel).
The developed software includes two full-featured models that demonstrate some of the basic functions of the Melos system. The first of these, Homunculus, implements the prototype of a newborn child, developing in an environment defined by the following contexts:
a) parental environment (family environment);
b) peers of early childhood (kindergarten and early playmates);
c) schoolmates (including other social contacts of adolescence);
d) colleagues at work (working contacts);
e) older friends (contacts with people who are not part of the professional environment).
In this case, each context becomes available either in a strictly allotted time, or when the subject reaches a certain level of development (expressed in possession of appropriate resources). In the process of simulated evolution, the subject realizes his needs, develops his own inclinations and preferences - under the influence of changing environmental conditions, as well as taking into account his own originally defined motivational attitudes. В результате у субъекта формируется собственный уникальный мотивационный портрет, управляющий профессиональным выбором на границе 20-25 лет, когда субъект выбирает один из двух введенных в модель вариантов самореализации (например: карьера ученого или карьера менеджера). Модель позволяет наблюдать удовлетворенность субъекта на любой стадии его развития, демонстрирует динамику изменения его мотивационной иерархии (возможен мониторинг по любому выбранному узлу). Также возможно изменение любых условий окружения или внутренних акцентуаций субъекта с целью произвольного модулирования и последующего анализа изменившихся поведенческих паттернов.
Вторая разработанная на системе «Melos» модель — Business Firm; она исследует влияние, которое оказывают на профессиональную продуктивность индивидуума его личностные предпочтения и склонности, напрямую не связанные с профессиональной деятельностью. В основу модели положена структура небольшого предприятия, состоящая из 8 человек, между которыми распределены 4 различные профессиональные роли. Между собой эти роли образуют функционально связанную иерархию, в которой от производительности верхнего узла зависит производительность подчиненных ему нижних узлов. Общая производительность предприятия зависит от адекватности и согласованности производственной отдачи всех уровней иерархии. Однако каждая из ролей может быть представлена несколькими работниками (например — два менеджера, четыре клерка). И при этом каждый работник характеризован своими собственными индивидуальными чертами (называемых мотивационным портретом, т.е., грубо говоря — тем комплексом, который побуждает данного человека к какой-либо деятельности вообще). Влияние этих особенностей на его продуктивность наглядно демонстрируется динамическими графиками, создающимися в процессе моделирования работы вышеописанного предприятия. Отдача каждого из работников отражает его личное отношение к тому роду деятельности, который он вынужден выполнять в соответствии с его профессиональными обязанностями. Модели подобного рода могут быть полезны социологам, HR-менеджерам и начальникам подразделений, которые желают распределять функциональные обязанности в соответствии с индивидуальными качествами их подчиненных.
На основании модели Business Firm создана презентация, в лаконичной форме демонстрирующая ее основные этапы, а также результаты моделирования.
Conclusion
Являясь технологическим ноу-хау, которое не имеет аналогов как на рынке прикладного ПО, так и среди академических разработок, система моделирования «Melos» находится в постоянной эволюции и совершенствовании. Ее возможности не ограничиваются указанным выше функциональным перечнем и спектром возможного применения. Зарекомендовав свой высокий потенциал в первых же экспериментах, инструмент моделирования может представлять интерес для коммерческих организаций и научно-исследовательских институтов в следующих формах:
а) реализация алгоритмов и построение engine-модулей в категории embedded-software;
б) создание программных комплексов категории социально-психологического инструментария;
в) продажа/передача технологии «Melos», исходных текстов и документации «in toto».
Кроме того, коллектив разработчиков будет рад любым предложениям сотрудничества, особенно, ориентированным на дальнейшее совершенствование системы.
Мы надеемся, все вышеизложенное послужит достаточным основанием для того, чтобы и вы заинтересовались нашей технологией, являющейся в настоящий момент уникальной и передовой среди аналогичных попыток построения Artificial Personality и моделирования поведения реальных индивидуумов.
Вы сможете узнать больше о системе моделирования «Melos» и созданных на ее основе прикладных программах, посетив наш сайт по адресу: или melos.info (англозычная страница) или melos.info/rus/ (русскаязычная страница), или запросив соответствующие материалы по email val@melos.info. Мы готовы ответить на любые вопросы, относящиеся к реализации, сфере применения и дальнейших перспективах развития системы.
Respectfully,
руководитель проекта «Melos»
Валентин Иванович Лохоня.
07.2001 — 04.2006
Continuing the theme:
Моделирование поведения человека при помощи новой технологии Искусственного Интеллекта — MELOS
Comments
To leave a comment
Mathematical Methods in Psychology
Terms: Mathematical Methods in Psychology