https://arxiv.org/abs/1809.06101
Invited talk at IRN Terascale, Annecy, France, 2019
https://indico.in2p3.fr/event/18701/contributions/72003/
Requirements: Python3 & Tensorflow 1.8+ & ROOT libraries
This is a research level proof-of-principle code. Depending on the physics application, additional algorithms, estimators and regularization techniques may be needed.
root printascii.c+ -b -q
train.sh
predict.sh
make && ./deeplot
If you use this work in your research, please cite the paper:
@article{mieskolainen2018deepefficiency,
title={DeepEfficiency - optimal efficiency inversion in higher dimensions at the LHC},
author={Mikael Mieskolainen},
year={2018},
journal={arXiv:1809.06101},
eprint={1809.06101},
archivePrefix={arXiv},
primaryClass={physics.data-an}
}