Techniques for batch reinforcement learning in robotics

This thesis addresses the Batch Reinforcement Learning methods in Robotics. This sub-class of Reinforcement Learning has shown promising results and has been the focus of recent research. Three contributions are proposed that aim to extend the state-of-art methods allowing for a faster and more stable learning process, such as required for learning in Robotics. The Q-learning update-rule is widely applied, since it allows to learn without the presence of a model of the environment. However, this update-rule is transition-based and does not take advantage of the underlying episodic structure of collected batch of interactions. The Q-Batch update-rule is proposed in this thesis, to process experiencies along the trajectories collected in the interaction phase. This allows a faster propagation of obtained rewards and penalties, resulting in faster and more robust learning. Non-parametric function approximations are explored, such as Gaussian Processes. This type of approximators allows to encode prior knowledge about the latent function, in the form of kernels, providing a higher level of exibility and accuracy. The application of Gaussian Processes in Batch Reinforcement Learning presented a higher performance in learning tasks than other function approximations used in the literature. Lastly, in order to extract more information from the experiences collected by the agent, model-learning techniques are incorporated to learn the system dynamics. In this way, it is possible to augment the set of collected experiences with experiences generated through planning using the learned models. Experiments were carried out mainly in simulation, with some tests carried out in a physical robotic platform. The obtained results show that the proposed approaches are able to outperform the classical Fitted Q Iteration.

O ensino do empreendedorismo: o papel do docente, os métodos pedagógicos e a sua relação com a intenção empreendedora

Can entrepreneurship be taught, or be learned? This is a question that brings a broad debate to the table. Indubitably, education has an essential role in attitude development and competences promotion. This assumption has fostered entrepreneurship initiatives and course proliferation, seeking to teach entrepreneurship to individuals. The increasing importance of the entrepreneurial spirit leads to the creation and promotion of entrepreneurship initiatives and encouragement of youngsters to have entrepreneurial profiles. Today, entrepreneurship education has a crucial role promoting entrepreneurial mindset s in younger people. It stimulates competence and skill development, which extends beyond the business world. The effects of such education depend on strategy and teaching pedagogies, but mostly on its effective implementation. This gives teachers and higher education institutions the challenge of finding nontraditional methodologies and alternative ways to teach the subject. Therefore, this study aims to explore, through the theory of planned behavior, the effects of entrepreneurship education and its relationship with entrepreneurial intention. Additionally, the essential and challenging role of the teacher is explored, contributing to a better understanding of its significance in entrepreneurship education that is remarkably poorly explored in the literature. Findings reinforce the strength of the theory of planned behavior as an intention measure; in addition, students were found to increase their entrepreneurial intentions, entrepreneurial knowledge and institutional context perceptions after taking a higher education course. Concerning teachers, a consensus was found between their education aims and similarities among their pedagogic methods, topics and forms of evaluation, along with time and resource management struggles.