Autonomous Robot Learning through Interaction with Humans
Prof. Dongheui Lee
Time: Tuesday May 24, 9:15 - 10:00
Imitation learning, one of the main streams for robot learning, provides an efficient way to learn new skills through human guidance, which can reduce time and cost to program the robot. In order to address autonomous motor skill learning and control in complex task scenarios, the presenter and her group have been working on a variety of fundamental sub-problems: movement primitive representation, reaction and adaptation in real world, the link between perception and action, and incremental learning. In this context, our mimesis model from partial observations provides a comprehensive framework for learning, recognition and reproduction of whole body motions even in case of incomplete measurement data. It allows several extensions and various applications, e.g., tool manipulation, 3D motion capturing with a monocular camera, reactive obstacle avoidance and adaptive robot collaboration. The inference mechanism can be applied not only to learn the robot's free body motion but also to learn physical interaction tasks, including human robot interaction. For autonomous systems incremental learning becomes an essential feature. During the talk I will focus on some related works including the refinement of learned skills via heterogeneous learning methods (e.g. by kinesthetic coaching), enhancement of human-robot cooperation tasks over time, and iterative learning control. The proposed algorithms are verified via implementations on a human skeleton model and multiple robotic systems including full size humanoid robots.
CV: Dongheui Lee is an assistant professor at the Institute of Automatic Control Engineering, Department of Electrical and Computer Engineering, Technical University of Munich, Germany since 2009. She is the head of Dynamic Human Robot Interaction for Automation System Lab and Carl-von-Linde Fellow at TUM Institute for Advanced Study. She received her B.S. and M.S degrees at the department of mechanical engineering, Kyunghee University, Korea, in 2001 and 2003, respectively. She worked as a research scientist at the Advanced Robotics Research Center, Korea Institute of Science and Technology (KIST) from 2001 to 2004. In 2007, she received her PhD degree at the department of Mechano‐Informatics, the University of Tokyo, Japan. After receiving PhD degree she joined the center of Information and Robot Technology at the University of Tokyo as a project assistant professor. Her research interests include human motion understanding, human robot interaction, machine learning in robotics, and mobile robot navigation. She is Coordinator of euRobotics Topic Group on physical Human Robot Interaction and the co-coordinator of TUM Center of Competence Robotics, Autonomy and Interaction.
How Social Biometrics and Cognitive Intelligence Changing the Way We Perceive Our World
Prof. Marina L. Gavrilova
Time: Tuesday May 24, 13:15 - 14:00
Our society continues to undergo tremendous growth with respect to all aspects of information access and sharing. It had a profound effect on the way we, humans, and the whole society lives, works and interacts in business and social settings. The terabytes of information being shared through social networks, on-line communities, games, software development tools, e-mails, blogs, posts, etc. is enormous. It also ranges in type: text, images, hyperlinks, likes, network connections, etc. What’s more, human social, behavioral and even cognitive traits are becoming more and more visible through interlinking of heterogeneous communications in on-line and off-line settings. This phenomenon gave rise to the rise of a new concept: Social Biometrics, that attempts to understand and extrapolate trends related to all aspects of human social interactions. The talk is devoted to definitions, examples, case studies and very recent research trends in this domain. The introduced concepts will be further illustrated through three case studies: cognitive multi-modal security system architecture, establishing identity of Twitter users through social networks analysis, and gender recognition of Flickr users based on human aesthetic preferences.
CV: Marina L. Gavrilova is a Full Professor in the Department of Computer Science, University of Calgary. Dr. Gavrilova’s research interests lie in the areas of biometric security, cognitive sciences, pattern recognition, social networks, and cyberworlds. Prof. Gavrilova is the founder and co-director of the Biometric Technologies Laboratory and SPARCS lab, with over 150 publications, including the World Scientific Bestseller (2007): Image Pattern Recognition: Synthesis and Analysis in Biometrics and IGI (2013) book Multimodal Biometrics and Intelligent Image Processing for Security System. She is a Founding Editor-in-Chief of Transactions on Computational Science journal, Springer, and an Associate Editor of the Visual Computer and the International Journal of Biometrics. Prof. Gavrilova has given invited lectures at leading international conferences (3AI, CW, WSCG, GRAPHICON, PSC, ICCI*CC, MIT, ICBAKE, etc), and appeared as guest at DIMACS Rutgers University/Bell Labs, USA; Microsoft Research, Redmond, USA; Samsung Research, South Korea; SERIAS Purdue University, USA, among other universities. Her research was profiled in newspaper and TV interviews, most recently featured in Exhibit at National Museum of Civilization, Quebec (2012), on Discovery Channel Canada Jay Ingram segment (2013) and in Business Magazine, Calgary, Alberta (2014).
Gamification with virtual characters at the borders of mixed realities and algebras
Assist. Prof. George Papagiannakis
Time: Wednesday May 25, 9:15 - 10:00
“Gamification” as the use of game design elements in non-game contexts, has attracted significant attention in various application fields of mixed reality, such as cultural heritage and education. At the same time, the rediscovery of geometric algebra via modern GPU-enhanced generators empowers computer graphics scientists with powerful new tools. These allow the replacement of different types of algebras, and their in-between conversions between CPU and GPU, such as linear algebra matrices, quaternions, dual-quaternions and Euler angles with a single geometric algebra mathematical representation. Such a representation enables enhanced gamified character simulation in different realities in the mixed reality spectrum, such as VR and AR. In this talk I will be presenting our latest results in these areas.
CV: George Papagiannakis is a computer scientist specialized in computer graphics and virtual-augmented reality. He obtained his PhD in Computer Science at the University of Geneva in Switzerland in 2006, his M.Sc. (Hons) in Advanced Computing at the University of Bristol in UK and his B.Eng. (Hons) in Computer Systems Engineering, at the University of Manchester, UK. Since 2011 he is assistant professor at the Computer Science department of the University of Crete, Greece and Research Fellow at the Computer Vision and Robotics Laboratory at the Institute of Computer Science of the Foundation for Research and Technology Hellas, Heraklion, Greece. Currently he is working on high-fidelity presence and human-computer interaction with virtual characters and humanoids in mixed reality. In 2011 he has been awarded with a Marie-Curie Intra-European Fellowship for Career Development from the European Commission’s Research Executive Agency. In 2015 he was awarded tenure from the Computer Science Department of the University of Crete, Greece.