Speaker
Adrian-Antonio Petre
(ETH Zurich (CH))
Description
In High-Energy Physics, generating meaningful parton configurations from a collision reconstructed within a detector is a critical step for many complex tasks like the Matrix Element Method computation and Bayesian inference on parameters of interest.
We propose to tackle this problem from a new perspective by using a Transformer network to analyze the full event at the reconstruction level (including jets and leptons). This approach extracts a latent vector which is used to condition a Flow network. The full architecture generates probable sets of partons that are compatible with the observed objects.
Our strategy is applicable to events with multiple jets multiplicity and can model additional radiation at parton level.
Authors
Adrian-Antonio Petre
(ETH Zurich (CH))
Davide Valsecchi
(ETH Zurich (CH))
Mauro Donega
(ETH Zurich (CH))
Rainer Wallny
(ETH Zurich (CH))
