Deep-learning jets with uncertainties and more
Sven Bollweg, Manuel Haussmann, Gregor Kasieczka, Michel Luchmann, Tilman Plehn, Jennifer Thompson
Abstract
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.
Topics & Concepts
Stability (learning theory)Artificial neural networkBayesian probabilityArtificial intelligenceComputer scienceMachine learningBayesian networkTraining setDeep learningAlgorithmProbabilistic logicControl (management)Jet (fluid)Energy (signal processing)Statistical modelDeep neural networksTracking (education)Structured predictionBayesian inferenceTrack (disk drive)Training (meteorology)Statistical learningTerm (time)PhysicsBayes' theoremBayesian statisticsDynamic Bayesian networkParticle physics theoretical and experimental studiesComputational Physics and Python ApplicationsAdversarial Robustness in Machine Learning