To create a realistic environment, many simulations require simulated agents with human behavior patterns. Manually creating such agents with realistic behavior is often a tedious and time-consuming task. This dissertation describes a new approach that automatically builds human behavior models for simulated agents by observing human performance. The research described in this dissertation synergistically combines Context-Based Reasoning, a paradigm especially developed to model tactical human performance within simulated agents, with Genetic Programming, a machine learning algorithm to construct the behavior knowledge in accordance to the paradigm. This synergistic combination of well-documented AI methodologies has resulted in a new algorithm that effectively and automatically builds simulated agents with human behavior. This algorithm was tested extensively with five different simulated agents created by observing the performance of five humans driving an automobile simulator. The agents show not only the ability/capability to automatically learn and generalize the behavior of the human observed, but they also capture some of the personal behavior patterns observed among the five humans. Furthermore, the agents exhibited a performance that was at least as good as agents developed manually by a knowledgeable engineer.

In this thesis, a new approach is developed, avoiding the correspondence problem caused by the difference in embodiment between imitator and demonstrator in imitation learning. In our work, the imitator is a fluidic system of dynamics totally different than the imitatee, which is a human performing hand gestures and human body postures. The fluidic system is composed of fluid particles, which are used for the discretization of the problem domain. In this work, we demonstrate the fluidics formation control so as to imitate by observation initially given human body poses and hand gestures. Our fluidic formation control is based on setting suitable parameters of Smoothed Particle Hydrodynamics (SPH), which is a particle based Lagrangian method, according to imitation learning. In the controller part, we developed three approaches: In the first one, we used Artificial Neural Networks (ANN) for training of the input-output pairs on the fluidic imitation system. We extracted shape based feature vectors for human hand gestures as inputs of the system and for output we took the fluid dynamics parameters. In the second approach, we employed the Principal Component Analysis (PCA) method for human hand gesture and human body pose classification and imitation. Lastly, we developed a region based controller which assigns the fluid parameters according to the human body poses and hand gestures. In this controller, our algorithm determines the best fitting ellipses on human body regions and human hand finger positions and maps ellipse parameters to the fluid parameters. The fluid parameters adjusted by the fluidics imitation controller are body force (f), density, stiffness coefficient and velocity of particles (V) so as to lead formations of fluidic swarms to human body poses and hand gestures.

Buscou-se investigar se a aprendizagem social pode facilitar a ocorrência de respostas supersticiosas, selecionadas pela relação acidental com eventos ambientais contíguos ao responder. Em uma atividade de computador, participantes podiam clicar em um mouse e havia um galão de água na tela que poderia encher-se de acordo com o esquema de reforçamento programado. O Experimento 1, realizado com oito estudantes universitários, foi composto por três sessões (delineamento ABA), alternando entre condições nas quais o participante atuou sozinho ou na presença de outro participante que também realizou a tarefa no mesmo momento. O segundo participante, porém, era um participante confederado com a experimentadora e respondeu à tarefa de forma previamente treinada. Os participantes foram expostos, em cada sessão, a esquemas de tempo variável de 15s (VT15s) e extinção (EXT) que foram alternados entre si a cada dois minutos. O participante confederado foi exposto aos esquemas de intervalo variável (VI15s) e EXT, e respondeu de forma constante em ambos os esquemas. Manipulou-se duas variáveis: o esquema de contiguidade e a exposição a outro indivíduo comportando-se na tarefa. Efeitos sobre as taxas de respostas dos participantes foram medidos. Não houve diferenciação do responder em função da mudança do esquema, porém houve diferença significativa entre as médias de respostas nas sessões individuais e conjuntas, permitindo a conclusão de que a manipulação da variável social produziu um efeito sobre as respostas. O Experimento 2 manipulou apenas a variável social. O esquema de reforçamento foi mantido em VT15s durante a sessão inteira em todo o experimento. Trabalhou-se com dois grupos de participantes, um no qual os participantes realizavam a sessão individualmente e outro grupo no qual outros participantes trabalhavam na sessão juntamente com um segundo participante, que, tal como no primeiro estudo, era um confederado que respondia de forma constante e previamente treinada, sendo exposto ao esquema de VI15s durante a sessão inteira. Observou-se uma diferença significativa entre as médias de respostas entre os dois grupos, o que mostra o efeito da variável manipulada. Além disso, observou-se uma variância significativamente maior das taxas de respostas dos participantes no grupo sem confederado em comparação com as dos participantes do grupo com confederado, que mostra um efeito de interação entre a variável social e o esquema de VT. O presente trabalho acrescenta dados novos tanto para a literatura de aprendizagem social quanto para a literatura de comportamento supersticioso, ao constatar que a variável social foi responsável por diminuir a variação das taxas de respostas de participantes expostos a um esquema independente, que, além disso foi independente desde o início da tarefa. Ademais confirmou-se achados sobre efeitos de facilitação social, sobretudo em uma situação de esquema não-contingente / The purpose of the present work was to investigate if social learning may facilitate occurrence of superstitious responses, selected by accidental relation with contiguous environmental events. In a computer activity, participants could click with a mouse and there was a water gallon on the screen that could fill itself according to reinforcement schedule programmed. The First Experiment was done with eight college students. It was composed by three sessions (ABA design), alternating between conditions in which the participant acted alone or in the presence of another participant, who also performed the task at the same moment. The second participant, however, was confederate with the experimenter and responded to the task in a previously trained manner. Participants were exposed, in each session, to 15s variable time schedules (VT15s) and extinction (EXT) which were alternated between each other every two minutes. The confederate participant was exposed to variable interval (VI15s) and exctinction (EXT) schedules and responded steadily in both schedules. Two variables were manipulated: the contiguity schedule and the exposure to another individual behaving on the task. Effects on participants rates of responding were measured. There was no differentiation in responding due to schedule change, however there was significant difference between mean rates of responses in individual sessions and in joint sessions, allowing the conclusion that the social variable manipulation had an effect on responses. The Second Experiment manipulated only the social variable. The schedule of reinforcement was maintained VT15s during the whole session in all the experiment. There were two groups of participants, one in which participants performed the session individually and another in which other participants performed in a session together with another participant. This participant, as in the First Experiment, was a confederate that responded steadily and in a previously trained manner, and was exposed to a VI15s schedule It was observed a significant difference between means of response of each group, which shows the effect of the manipulated variable. Moreover there was a significant higher variance in response rates of the individual group participants in comparison with response rate of the group with a confederate, which shows an effect of interaction between the social variable and the VT schedule. The present work adds new data to social learning literature as much as to superstitious behavior literature, upon verifying that the social variable was responsible for diminishing variation in response rates of participants exposed to an independent schedule, which beyond that, was independent since the beginning of the task. Furthermore it confirmed some findings of social facilitation effects, especially in a situation of non-contingent schedule

Buscou-se investigar se a aprendizagem social pode facilitar a ocorrência de respostas supersticiosas, selecionadas pela relação acidental com eventos ambientais contíguos ao responder. Em uma atividade de computador, participantes podiam clicar em um mouse e havia um galão de água na tela que poderia encher-se de acordo com o esquema de reforçamento programado. O Experimento 1, realizado com oito estudantes universitários, foi composto por três sessões (delineamento ABA), alternando entre condições nas quais o participante atuou sozinho ou na presença de outro participante que também realizou a tarefa no mesmo momento. O segundo participante, porém, era um participante confederado com a experimentadora e respondeu à tarefa de forma previamente treinada. Os participantes foram expostos, em cada sessão, a esquemas de tempo variável de 15s (VT15s) e extinção (EXT) que foram alternados entre si a cada dois minutos. O participante confederado foi exposto aos esquemas de intervalo variável (VI15s) e EXT, e respondeu de forma constante em ambos os esquemas. Manipulou-se duas variáveis: o esquema de contiguidade e a exposição a outro indivíduo comportando-se na tarefa. Efeitos sobre as taxas de respostas dos participantes foram medidos. Não houve diferenciação do responder em função da mudança do esquema, porém houve diferença significativa entre as médias de respostas nas sessões individuais e conjuntas, permitindo a conclusão de que a manipulação da variável social produziu um efeito sobre as respostas. O Experimento 2 manipulou apenas a variável social. O esquema de reforçamento foi mantido em VT15s durante a sessão inteira em todo o experimento. Trabalhou-se com dois grupos de participantes, um no qual os participantes realizavam a sessão individualmente e outro grupo no qual outros participantes trabalhavam na sessão juntamente com um segundo participante, que, tal como no primeiro estudo, era um confederado que respondia de forma constante e previamente treinada, sendo exposto ao esquema de VI15s durante a sessão inteira. Observou-se uma diferença significativa entre as médias de respostas entre os dois grupos, o que mostra o efeito da variável manipulada. Além disso, observou-se uma variância significativamente maior das taxas de respostas dos participantes no grupo sem confederado em comparação com as dos participantes do grupo com confederado, que mostra um efeito de interação entre a variável social e o esquema de VT. O presente trabalho acrescenta dados novos tanto para a literatura de aprendizagem social quanto para a literatura de comportamento supersticioso, ao constatar que a variável social foi responsável por diminuir a variação das taxas de respostas de participantes expostos a um esquema independente, que, além disso foi independente desde o início da tarefa. Ademais confirmou-se achados sobre efeitos de facilitação social, sobretudo em uma situação de esquema não-contingente / The purpose of the present work was to investigate if social learning may facilitate occurrence of superstitious responses, selected by accidental relation with contiguous environmental events. In a computer activity, participants could click with a mouse and there was a water gallon on the screen that could fill itself according to reinforcement schedule programmed. The First Experiment was done with eight college students. It was composed by three sessions (ABA design), alternating between conditions in which the participant acted alone or in the presence of another participant, who also performed the task at the same moment. The second participant, however, was confederate with the experimenter and responded to the task in a previously trained manner. Participants were exposed, in each session, to 15s variable time schedules (VT15s) and extinction (EXT) which were alternated between each other every two minutes. The confederate participant was exposed to variable interval (VI15s) and exctinction (EXT) schedules and responded steadily in both schedules. Two variables were manipulated: the contiguity schedule and the exposure to another individual behaving on the task. Effects on participants rates of responding were measured. There was no differentiation in responding due to schedule change, however there was significant difference between mean rates of responses in individual sessions and in joint sessions, allowing the conclusion that the social variable manipulation had an effect on responses. The Second Experiment manipulated only the social variable. The schedule of reinforcement was maintained VT15s during the whole session in all the experiment. There were two groups of participants, one in which participants performed the session individually and another in which other participants performed in a session together with another participant. This participant, as in the First Experiment, was a confederate that responded steadily and in a previously trained manner, and was exposed to a VI15s schedule It was observed a significant difference between means of response of each group, which shows the effect of the manipulated variable. Moreover there was a significant higher variance in response rates of the individual group participants in comparison with response rate of the group with a confederate, which shows an effect of interaction between the social variable and the VT schedule. The present work adds new data to social learning literature as much as to superstitious behavior literature, upon verifying that the social variable was responsible for diminishing variation in response rates of participants exposed to an independent schedule, which beyond that, was independent since the beginning of the task. Furthermore it confirmed some findings of social facilitation effects, especially in a situation of non-contingent schedule