User menu

Accès à distance ? S'identifier sur le proxy UCLouvain

MadNIS -- Neural Multi-Channel Importance Sampling

Primary tabs

download
  • Open access
  • PDF
  • 869.83 K
Winterhalder, Ramon [UCL] Maltoni, Fabio [UCL] Mattelaer, Olivier [UCL]

Theory predictions for the LHC require precise numerical phase-space integration and generation of unweighted events. We combine machine-learned multi-channel weights with a normalizing flow for importance sampling, to improve classical methods for numerical integration. We develop an efficient bi-directional setup based on an invertible network, combining online and buffered training for potentially expensive integrands. We illustrate our method for the Drell-Yan process with an additional narrow resonance.

Document type Article de périodique (Journal article) – Article de recherche
Access type Accès libre
Publication date 2023
Language Anglais
Journal information "SciPost Physics" - Vol. 15, art. 141 (2023)
Peer reviewed yes
issn 2542-4653
Publication status Publié
Affiliation UCL - SST/IRMP - Institut de recherche en mathématique et physique
Keywords network ; flow ; interpretation of experiments: CERN LHC Coll ; narrow resonance ; phase space ; on-line ; Drell-Yan process
Links
Bibliographic reference Winterhalder, Ramon ; Maltoni, Fabio ; Mattelaer, Olivier ; et. al. MadNIS -- Neural Multi-Channel Importance Sampling. In: SciPost Physics, Vol. 15, art. 141 (2023)
Permanent URL http://hdl.handle.net/2078.1/278652