SciPost Submission Page

Understanding Event-Generation Networks via Uncertainties

by Marco Bellagente, Manuel Haußmann, Michel Luchmann, Tilman Plehn

Submission summary

As Contributors:	Marco Bellagente · Tilman Plehn
Arxiv Link:	https://arxiv.org/abs/2104.04543v2 (pdf)
Date accepted:	2021-10-08
Date submitted:	2021-10-04 08:52
Submitted by:	Bellagente, Marco
Submitted to:	SciPost Physics
Academic field:	Physics
Specialties:	High-Energy Physics - Phenomenology
Approaches:	Computational, Phenomenological

Abstract

Following 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.

Published as SciPost Phys. 13, 003 (2022)

---