MadNIS -- Neural Multi-Channel Importance Sampling
High Energy Physics - Phenomenology
2023-10-04 v2 High Energy Physics - Experiment
Computational Physics
Abstract
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.
Keywords
Cite
@article{arxiv.2212.06172,
title = {MadNIS -- Neural Multi-Channel Importance Sampling},
author = {Theo Heimel and Ramon Winterhalder and Anja Butter and Joshua Isaacson and Claudius Krause and Fabio Maltoni and Olivier Mattelaer and Tilman Plehn},
journal= {arXiv preprint arXiv:2212.06172},
year = {2023}
}
Comments
33 pages, 15 figures, minor fixes to v1