NBIA Seminar: Philipp Eller

The landscape of experimental neutrino detectors and telescopes is continuously expanding, and the amounts of quality data is ever growing. In this talk I highlight a few examples of our research where we borrow, develop, or apply novel statistical techniques to extract more information out of our data than previously possible. The applications shown will include unsupervised learning for astrophysical neutrino flare detection, Cherenkov detector event reconstruction techniques, advancements in MCMC sampling for parameter inference, non-parametric tests for supernovae triggering, and fast approximations of physics models using deep neural networks.