Speaker: Dr. Sven Krippendorf of LMU Munich

Abstract: In this colloquium I will report on how we can gain valuable insights on key physics questions by utilizing Machine Learning (ML). I will report on how ML helps our search for dark matter (in particular axion-like particles), how ML helps us to understand phenomenological signals of quantum gravity, and how we can use ML to discover mathematical and physical structures such as symmetries and dualities. I will discuss how the latter helps us in finding interpretable features in neural networks and how our physicists’ bias can help us to select good neural network architectures.