Papers

Solving Simulation Systematics in and with AI/ML

B Viren, J Huang, Y Huang, M Lin, Y Ren… - arXiv preprint arXiv …, 2022 - arxiv.org
Physics paper hep-ex Suggest

… A broad class architectures have been collected under the general umbrella of “simulation-based inference” [9]. In addition, a trainable simulation can provide a potentially powerful …

Link to paper

BibTeX

@article{2203.06112v1,
Author        = {Brett Viren and Jin Huang and Yi Huang and Meifeng Lin and Yihui Ren and Kazuhiro Terao and Dmitrii Torbunov and Haiwang Yu},
Title         = {Solving Simulation Systematics in and with AI/ML},
Eprint        = {2203.06112v1},
ArchivePrefix = {arXiv},
PrimaryClass  = {hep-ex},
Abstract      = {Training an AI/ML system on simulated data while using that system to infer
on data from real detectors introduces a systematic error which is difficult to
estimate and in many analyses is simply not confronted. It is crucial to
minimize and to quantitatively estimate the uncertainties in such analysis and
do so with a precision and accuracy that matches those that AI/ML techniques
bring. Here we highlight the need to confront this class of systematic error,
discuss conventional ways to estimate it and describe ways to quantify and to
minimize the uncertainty using methods which are themselves based on the power
of AI/ML. We also describe methods to introduce a simulation into an AI/ML
network to allow for training of its semantically meaningful parameters. This
whitepaper is a contribution to the Computational Frontier of Snowmass21.},
Year          = {2022},
Month         = {Mar},
Url           = {http://arxiv.org/abs/2203.06112v1},
File          = {2203.06112v1.pdf}
}

Share