Understanding Event-Generation Networks via Uncertainties
2022; SciPost.org; Volume: 13; Issue: 1 Linguagem: Inglês
10.21468/scipostphys.13.1.003
ISSN2542-4653
AutoresMarco Bellagente, Manuel Haußmann, Michel Luchmann, Tilman Plehn,
Tópico(s)High-Energy Particle Collisions Research
ResumoFollowing the growing success of generative neural networks in LHC simulations, the crucial question is how to control the networks and assign uncertainties to their event output. We show how Bayesian normalizing flow or invertible networks capture uncertainties from the training and turn them into an uncertainty on the event weight. Fundamentally, the interplay between density and uncertainty estimates indicates that these networks learn functions in analogy to parameter fits rather than binned event counts.
Referência(s)