Deep Reinforcement Learning in Robotics Use CasesNuno Lau, Universidade de Aveiro, PortugalAvailable SoonMichele Magno, ETH Zurich, Switzerland
Short Bio Nuno Lau is Associate Professor at Aveiro University, Portugal. He got is Electrical Engineering Degree from Oporto University in 1993, a DEA degree in Biomedical Engineering from Claude Bernard University, Lyon, in 1994 and the PhD from Aveiro University in 2003. His research interests include Intelligent Robotics, Artificial Intelligence, Multi-Agent Systems and Simulation. Nuno Lau has participated, often with the coordination role, in several research projects that have been awarded international prizes. His participation and leading role in RoboCup teams made him RoboCup World Champion 3 times in different Leagues and many other podium positions. Nuno Lau is the author of more than 150 publications in international conferences and journals, including several Best Paper awards. He Supervised 9 PhD students that concluded their degree, and is currently supervising 6 PhD students. Nuno Lau was President and Vice-President of the Portuguese Robotics Society. He is the Vice-Coordinator of the Institute of Electronics and Informatics Engineering of Aveiro, and Principal Investigator of the Intelligent Robotics and Systems group at the same research unit.
Abstract Deep Reinforcement Learning techniques are now widely used in many scientific disciplines. Using Deep Reinforcement Learning in Robotics is a very challenging task as robots may be expensive and fragile, and the time and effort to collect data is, in general, quite high. Considering these premises, we have developed several techniques that through the use of simulators, and adapted learning/optimization algorithms, that use the data very efficiently, make the use of Deep Reinforcement Learning in Robotics an effective option over hand-coded approaches and that may achieve impressive results. These techniques have been applied to solve perception, reasoning, behavior level and long horizon robotic tasks. This talk will present some of these techniques, in different contexts, namely those related to optimization/ML based development of robotic skills, Q-Batch update rule, multi-agent skill and communication learning, adapted interfaces, imperfect symmetry exploitation and long horizon skill sequencing and learning.
Short Bio Available Soon
Abstract Available Soon