Karlsruhe 2024 – scientific programme
Parts | Days | Selection | Search | Updates | Downloads | Help
T: Fachverband Teilchenphysik
T 67: Data, AI, Computing 5 (normalising flows)
T 67.7: Talk
Wednesday, March 6, 2024, 17:30–17:45, Geb. 30.33: MTI
Conditional normalizing flows for correcting simulations — •Caio Cesar Daumann1, Mauro Donega2, Johannes Erdmann1, Massimiliano Galli2, Jan Lukas Späh1, and Davide Valsecchi2 — 1III. Physikalisches Institut A, RWTH Aachen University — 2ETH Zurich
Simulated events are key ingredients for almost all high-energy physics analyses. However, imperfections in the configuration of the simulation often result in mis-modelling and lead to discrepancies between data and simulation. Such mis-modelling often must be taken into account by correction factors accompanied by systematical uncertainties, which can compromise the sensitivity of measurements and searches. To address this issue, we propose to use normalizing flows, a powerful technique for learning the underlying distributions of input data. Where the flow is trained with both simulation and data, and is able to map betwenn the distributions.
We demonstrate that the proposed architecture can accurately correct simulation marginal distributions results in better closure with data. Additionally, we show that the flow can correct the correlations of variables, leading to significantly reduced differences in correlation matrices compared to the data.
Keywords: High energy physics; Electrons; Normalizing flows; Higgs Boson