Short CV
Alexandra Kirsch received her diploma in computer science from TUM in 2003. She worked as a research and teaching assistant at TUM between 2003 and 2007 and received her doctoral degree. Afterwards, she worked as a management consultant at Booz & Co. In 2008 she returned to TUM as a senior research scientist at the Intelligent Autonomous Systems Group in the Faculty of Informatics and a leader of the independent Junior Research Group “Planning for Adaptive Robot Assistance” in the Excellence Cluster Cognition for Technical Systems (CoTeSys). Since April 2012 Alexandra Kirsch is an assistant professor at the Media Informatics Department of the University of Tübingen.
Selected Awards
- 2012, Member of the Bayerische Akademie der Wissenschaften Förderkolleg
- 2003, Award by Comet Computer GmbH for excellent graduation results
Research Interests
Alexandra Kirsch investigates control mechanisms for autonomous robots based on methods from Artificial Intelligence with the aim to develop flexible, adaptive robots, whose behavior is understandable for humans and allows a close collaboration between humans and robots. The central idea in her research is that intelligent robots need different kinds of knowledge about the world, in particular models about how the world evolves, how the robot's actions change the world, and how humans usually act. For acquiring and updating such models while the robot is working, Alexandra Kirsch has developed the Robot Learning Language (RoLL), which allows a robot to acquire experience and execute learning problems automatically. She explores how these models can be used in plan-based robot control for action selection, failure detection and adaptation to human needs. In the context of interaction methods with humans, she collaborates with psychologists and neuroscientists for model acquisition and the execution of joint human-robot planning tasks.
Selected Publications
- Kirsch, Alexandra: Robot learning language — Integrating programming and learning for cognitive systems. Robotics and Autonomous Systems 57 (9), 2009, 943-954.
- Kirsch, A.; Kruse, T.; Mösenlechner, L.: An integrated planning and learning framework for human-robot interaction. 4th Workshop on Planning and Plan Execution for Real-World Systems (held in conjuction with ICAPS 09), 2009.
- Kirsch, A. (Ed.): Integration of Programming and Learning in a Control Language for Autonomous Robots Performing Everyday Activities. Technische Universität München, 2008.
- Müller, A.; Kirsch, A.; Beetz, M.: Transformational planning for everyday activity. Proceedings of the 17th International Conference on Automated Planning and Scheduling (ICAPS'07), 2007.
- Kirsch, A.; Schweitzer, M.; Beetz, M.: Making robot learning controllable: A case study in robot navigation. Proceedings of the ICAPS Workshop on Plan Execution: A Reality Check, 2005.
- Müller, A.; Kirsch, A.; Beetz, M.: Object-orientated model-based extensions of robot control languages. 27th German Conference on Artificial Intelligence, 2004.
- Müller, A.; Kirsch, A.; Beetz, M.: RPL-LEARN: Extending an autonomous robot control language to perform experience-based learning. 3rd International Joint Conference on Autonomous Agents & Multi Agent Systems (AAMAS), 2004.
- Kirsch, Alexandra; Beetz, Michael: Combining Learning and Programming for High-Performance Robot Controllers. In: Autonome Mobile Systeme 2005. Springer Science + Business Media.
- Kirsch, Alexandra; Beetz, Michael: Training on the Job — Collecting Experience with Hierarchical Hybrid Automata. In: Lecture Notes in Computer Science. Springer Science + Business Media.
Publications as TUM-IAS-Fellow