1 resultado para learning in projects
em Repositório Institucional da Universidade de Aveiro - Portugal
Filtro por publicador
- Aberdeen University (2)
- Abertay Research Collections - Abertay University’s repository (1)
- Academic Archive On-line (Jönköping University; Sweden) (1)
- Academic Archive On-line (Karlstad University; Sweden) (2)
- Academic Research Repository at Institute of Developing Economies (1)
- Acceda, el repositorio institucional de la Universidad de Las Palmas de Gran Canaria. España (1)
- Adam Mickiewicz University Repository (1)
- AMS Tesi di Dottorato - Alm@DL - Università di Bologna (3)
- AMS Tesi di Laurea - Alm@DL - Università di Bologna (6)
- Archive of European Integration (7)
- Aston University Research Archive (58)
- Biblioteca Digital da Produção Intelectual da Universidade de São Paulo (4)
- Biblioteca Digital da Produção Intelectual da Universidade de São Paulo (BDPI/USP) (9)
- Biodiversity Heritage Library, United States (1)
- BORIS: Bern Open Repository and Information System - Berna - Suiça (39)
- Brock University, Canada (37)
- Bulgarian Digital Mathematics Library at IMI-BAS (6)
- CentAUR: Central Archive University of Reading - UK (52)
- Cochin University of Science & Technology (CUSAT), India (3)
- Comissão Econômica para a América Latina e o Caribe (CEPAL) (3)
- Consorci de Serveis Universitaris de Catalunya (CSUC), Spain (36)
- Corvinus Research Archive - The institutional repository for the Corvinus University of Budapest (2)
- CUNY Academic Works (1)
- Dalarna University College Electronic Archive (7)
- Digital Commons - Michigan Tech (2)
- Digital Commons @ DU | University of Denver Research (3)
- Digital Commons @ Winthrop University (1)
- Digital Commons at Florida International University (15)
- Digital Peer Publishing (7)
- DigitalCommons - The University of Maine Research (1)
- DigitalCommons@The Texas Medical Center (5)
- DigitalCommons@University of Nebraska - Lincoln (7)
- Doria (National Library of Finland DSpace Services) - National Library of Finland, Finland (47)
- DRUM (Digital Repository at the University of Maryland) (2)
- Escola Superior de Educação de Paula Frassinetti (2)
- Fachlicher Dokumentenserver Paedagogik/Erziehungswissenschaften (4)
- Glasgow Theses Service (2)
- Illinois Digital Environment for Access to Learning and Scholarship Repository (2)
- Institutional Repository of Leibniz University Hannover (1)
- Instituto Politécnico de Viseu (2)
- Instituto Politécnico do Porto, Portugal (35)
- Martin Luther Universitat Halle Wittenberg, Germany (2)
- Massachusetts Institute of Technology (1)
- Memorial University Research Repository (1)
- Ministerio de Cultura, Spain (19)
- National Center for Biotechnology Information - NCBI (8)
- Nottingham eTheses (1)
- Open Access Repository of Association for Learning Technology (ALT) (8)
- Open University Netherlands (2)
- Portal de Revistas Científicas Complutenses - Espanha (3)
- Portal do Conhecimento - Ministerio do Ensino Superior Ciencia e Inovacao, Cape Verde (1)
- Publishing Network for Geoscientific & Environmental Data (1)
- QUB Research Portal - Research Directory and Institutional Repository for Queen's University Belfast (7)
- Repositório Aberto da Universidade Aberta de Portugal (3)
- Repositorio Académico de la Universidad Nacional de Costa Rica (1)
- Repositório Científico da Escola Superior de Enfermagem de Coimbra (1)
- Repositório Científico da Universidade de Évora - Portugal (2)
- Repositório Científico do Instituto Politécnico de Lisboa - Portugal (6)
- Repositório da Produção Científica e Intelectual da Unicamp (1)
- Repositório digital da Fundação Getúlio Vargas - FGV (2)
- Repositório Institucional da Universidade de Aveiro - Portugal (1)
- Repositório Institucional UNESP - Universidade Estadual Paulista "Julio de Mesquita Filho" (10)
- RUN (Repositório da Universidade Nova de Lisboa) - FCT (Faculdade de Cienecias e Technologia), Universidade Nova de Lisboa (UNL), Portugal (11)
- Savoirs UdeS : plateforme de diffusion de la production intellectuelle de l’Université de Sherbrooke - Canada (2)
- School of Medicine, Washington University, United States (1)
- Scielo Saúde Pública - SP (11)
- Scottish Institute for Research in Economics (SIRE) (SIRE), United Kingdom (4)
- Universidad de Alicante (4)
- Universidad del Rosario, Colombia (4)
- Universidad Politécnica de Madrid (11)
- Universidade Complutense de Madrid (2)
- Universidade do Minho (2)
- Universidade dos Açores - Portugal (2)
- Universitat de Girona, Spain (3)
- Universitätsbibliothek Kassel, Universität Kassel, Germany (4)
- Université de Lausanne, Switzerland (34)
- Université de Montréal, Canada (5)
- University of Canberra Research Repository - Australia (2)
- University of Connecticut - USA (2)
- University of Michigan (9)
- University of Queensland eSpace - Australia (68)
- University of Southampton, United Kingdom (4)
- University of Washington (4)
- WestminsterResearch - UK (2)
- Worcester Research and Publications - Worcester Research and Publications - UK (3)
Resumo:
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.