Invited talk by Dr. Lisa Gutzeit

Short abstract: Implementing new behaviors for robotic systems is tedious work. In recent years more and more often machine learning has been used to simplify this problem. By using learning from demonstration techniques, intuitive knowledge from humans is leveraged to initialize a refinement process that has to be done on the real system. In this talk, all important aspects of learning from demonstration, ranging from human demonstration and embodiment mapping to robotic movement refinement are explained. Additionally, a modular framework for transferring human movements to a robotic system is introduced along with evaluation results of this framework for different movements and robotic systems.

Short bio: Lisa Gutzeit received the bachelor’s and master’s degrees in computational life science from the University of Lübeck, Germany, in 2009 and 2011, respectively. In 2023 she received the Dr. rer. nat. degree from the University of Bremen, Germany. Since 2012, she has been working as a Researcher at the Robotics Research Group, University of Bremen, and since 2023 she is Senior Researcher at the Robotics Innovation Center, DFKI, Germany, where she co-leads the team Interaction & Learning. Her scientific research interests include the fields of machine learning and human–machine interaction.