Prof. Carme Torras
Manipulation Challenges in Assistive Robotics: Research and Ethics
Assistive robotics is a fast growing field aimed at helping caregivers in hospitals, rehabilitation centers and nursery homes, as well as empowering people with reduced mobility at home, so that they can live autonomously. Most tasks assistive robots need to perform (e.g., helping users to dress, guiding rehabilitation, feeding) require dexterous manipulation skills, which need to be easily taught and customized by non-experts. In addition, such skills must be very compliant and intrinsically safe to people, as well as able to deal with deformable materials like clothing. Some results of projects addressing these demanding challenges, such as CLOTHILDE, I-DRESS and SOCRATES, will be shown.
Assistive robots raise also fundamental ethical issues, many practical ones stemming from autonomous robot decision-making conflicting with human freedom and dignity. Several institutions are developing regulations and standards, and many ethics education initiatives include contents on human-robot interaction and human dignity in assistive situations. In the talk, educational materials from a university course on Ethics in Social Robotics and AI focusing on the assistive context will be presented.
Carme Torras (www.iri.upc.edu/people/torras) is Research Professor at the Robotics Institute (CSIC-UPC) in Barcelona, where she leads a research group on assistive and collaborative robotics. She received M.Sc. degrees in Mathematics and Computer Science from the University of Barcelona and the University of Massachusetts, respectively, and a Ph.D. degree in Computer Science from the Technical University of Catalonia (UPC). She has led sixteen European projects, the latest being her ERC Advanced Grant project CLOTHILDE – Cloth manipulation learning from demonstrations.
Prof. Torras is IEEE and EurAI Fellow, member of Academia Europaea and the Royal Academy of Sciences and Arts of Barcelona. She has served as Senior Editor of the IEEE Transactions on Robotics, Associate Vice-President for Publications of the IEEE Robotics and Automation Society (RAS), and member of the Administrative Committee of IEEE RAS in the period 2016-2018.
Convinced that science fiction can help promote ethics in robotics and new technologies, one of her novels – winner of the Pedrolo and Ictineu awards – has been translated into English with the title The Vestigial Heart (MIT Press, 2018) and published together with online materials to teach a course on “Ethics in Social Robotics and AI”.
Prof. Michiel van de Panne
Learned Agility for Robots and Digital Creatures
How is agile motion achieved in nature? How can we develop similar capabilities for robots and simulated creatures?
These questions have been of longstanding interest in robotics, computer graphics, control, biomechanics, kinesiology, neuromotor control, and machine learning.
Deep reinforcement learning (DRL) offers a potentially promising approach for learning agile behaviors.
However, there are ample drawbacks to overcome: reward functions can be unintuitive, simulation results may not transfer to the real world, solutions are sensitive to hyperparameter choices, can be expensive to compute, and are may not be interpretable.
In this talk, we review the most common DRL-based methods, relate these to common existing approaches control, and present recent agile motion results using DRL that originate from a number of research groups.
Results include simulated human and animal skills that can reproduce a large variety of highly dynamic motions, with varying degrees of biomechanical fidelity.
We further describe recent success in sim-to-real for the Cassie biped and other robots.
Lastly, we touch on some of the many open problems.
Michiel van de Panne is a Professor in the Department of Computer Science at UBC, with research interests that span reinforcement learning, robotics, motor learning, and physics-based animation. His group models robot, human, and animal movements and the motor skills that underly their movement. His research has been applied to robotics, games, and visual effects for film, and has been featured in the MIT Technology Review, Engadget, and elsewhere. His former students and postdocs have gone on to co-found startups (Element AI, Anomotion, VGC); have assumed key leadership related to AI, including at Tesla (Director of AI) and DeepMotion (Chief Scientist); and hold faculty positions at universities including ETH Zurich, Simon Fraser University, Leeds, and York University.