24–28 Aug 2026
Kirchhoff Institute for Physics (KIP)
Europe/Berlin timezone

Know What You Don't Flow

25 Aug 2026, 12:20
8m
HS1

HS1

Inference & Uncertainty 🔀 Inference & Uncertainty

Speaker

Lorenz Vogel (Institute for Theoretical Physics (ITP), Heidelberg University, Germany)

Description

Calibrated learned uncertainties are a key requirement also for generative neural networks in LHC physics. For a toy model with an explicit likelihood we show how a heteroscedastic and a Bayesian normalizing flow learn the systematic and statistical uncertainties on the underlying phase space density. Without an explicit likelihood we train the heteroscedastic loss on a classifier-reweighted approximate generative network. We illustrate our comprehensive approach for top pair events and show how a conditional heteroscedastic flow propagates calibrated uncertainties to all phase space directions.

Authors

Anja Butter (LPNHE, Sorbonne Université, Université Paris Cité, CNRS/IN2P3, Paris, France, Institut für Theoretische Physik, Universität Heidelberg, Germany) Sascha Diefenbacher (ITP Heidelberg) Tilman Plehn (Institut für Theoretische Physik, Universität Heidelberg; Interdisciplinary Center for Scientific Computing (IWR), Universität Heidelberg) Lorenz Vogel (Institute for Theoretical Physics (ITP), Heidelberg University, Germany)

Presentation materials