Decision-making under uncertainty: including prior knowledge from physics

• Applicant: Dr. Michael Muehlebach
• Subject Area: Automation, Mechatronics, Control Systems, Intelligent Technical Systems, Robotics
• Term: since 2021
• Project identifier: Deutsche Forschungsgemeinschaft (DFG) - Project number 456587626

The current paradigm in supervised machine learning and artificial intelligence is to fit high-capacity models to large amounts of data. Prior information, such as causal structures, first-principles, or a-priori known symmetries, is often discarded, and as a result, the predictions may not generalize to unforeseen situations, i.e., situations that are not adequately captured with the training data. We propose to analyze the problem of decision-making in such situations, which includes investigating and developing data-driven methods that explicitly incorporate prior knowledge from physics. The incorporation of prior information might also lead to an efficient quantification of prediction uncertainties and can potentially be exploited by optimization algorithms for fast decision-making. As a result, the research contributes to a safe and reliable deployment of machine learning and artificial intelligence in cyber-physical applications.

• DFG Programme: Independent Junior Research Groups