Speaker: Klemens Böhm, Karlsruhe Institute of Technology
The DFG-funded research-training group (RTG) on energy-status data does research on data-science methods that might help to arrive at better future energy systems and ultimately at a better energy footprint. In order to provide more details on this vision, my talk will cover three different research activities which have been going on/are currently ongoing within the RTG, at different levels of detail. These are the detection of changes in correlation in data streams, predictions from data streams in the presence of constraints, and active learning for robust distribution grids. In all three cases, familiarity with the domain is key, and I will explain the relationship to theory-guided data science, as I see it, as well.