The ELLIS unit Darmstadt will lay the foundations for intelligent systems that understand context and meaning, and are able to adapt accordingly. The aim is to make computers learn as much about the world, so rapidly and flexibly, as humans. To this end, the research activities of the unit comprise problem areas such as deep (probabilistic) models, hybrid models, physics-aware machine and robot learning, Bayesian (nonparametric) models, (deep and/or inverse) reinforcement learning, probabilistic programming, vision/NLP/robotics, explainable and interactive AI, and human-centric AI. The unit is part of a strategic high-level initiative of TU Darmstadt called Intelligent Systems at TU Darmstadt (IS•DA) which broadly connects the local activities in the areas of artificial intelligence (AI).

Unit Director