Michel Luchmann (Universität Heidelberg)
Bayesian neural networks allow us to keep track of uncertainties, for example in top tagging, by learning a tagger output together with an error band. We illustrate the main features of Bayesian versions of established deep-learning taggers. We show how they capture statistical uncertainties from finite training samples, systematics related to the jet energy scale, and stability issues through pile-up. Altogether, Bayesian networks offer many new handles to understand and control deep learning at the LHC without introducing a visible prior effect and without compromising the network performance.
Michel Luchmann (Universität Heidelberg) Tilman Plehn Gregor Kasieczka (Hamburg University (DE)) Jennifer Thompson (ITP Heidelberg) Sven Bollweg (Universität Hamburg) Manuel Haussmann (Universität Heidelberg)