We had a successful virtual spring school – with around 50 to 100 attendees for the five tutorial type sessions and the two evening talks. The students enjoyed the discussion with the lecturers and the broad perspectives on diverse topics (for detailed overview see the Spring School homepage) in Explainable Artificial Intelligence – ranging from an introduction to Explainable AI (Gregoire Montavon) over Symbolic and Sub-symbolic Representations (Maribel Acostato) to the Interpretability of Gaussian Processes (from our own Markus Lange-Hegermann). Furthermore, current hot topics as AutoML (Bernd Bischl) or Graph Neural Networks (Christopher Morris) were well covered. The two evening lectures, first, introduced the recent success in Deep Reinforcement Learning (a hands-on talk by Kaushik Subramanian on their nature paper Outracing champion Gran Turismo drivers with deep reinforcement learning) and, secondly, provided a great introduction to societal impact of AI through the talk by Rainer Mühlhoff (on the Ethics of AI).
Many of the students presented their own projects in the virtual gather.town poster session and two were awarded a poster prize: Congratulations (again) to: Annika Mütze, Matthias Rottmann, and Hanno Gottschalk (for their poster on A Downstream Task Informed Domain Adaptation GAN for Semantic Street Scene Segmentation) and to Federico Pennino, Shamini Koravuna, Christoph Ostrau, and Ulrich Rückert (N-MNIST object recognition with Spiking Neural Networks).