https://doi.org/10.1140/epjc/s10052-022-10070-0
Regular Article - Experimental Physics
Punzi-loss:
a non-differentiable metric approximation for sensitivity optimisation in the search for new particles
1
Aix Marseille Université, CNRS/IN2P3, CPPM, 13288, Marseille, France
2
Deutsches Elektronen-Synchrotron, Hamburg, Germany
3
Dipartimento di Fisica, Università di Pisa, 56127, Pisa, Italy
4
Dipartimento di Scienze Fisiche, Università di Napoli Federico II, 80126, Naples, Italy
5
Duke University, Durham, USA
6
Faculty of Mathematics and Physics, Charles University, Prague, Czech Republic
7
Helmholtz AI, Karlsruhe Institute of Technology, 76131, Karlsruhe, Germany
8
High Energy Accelerator Research Organization (KEK), Tsukuba, Japan
9
INFN-Sezione di Napoli, 80126, Naples, Italy
10
INFN-Sezione di Padova, Padua, Italy
11
INFN-Sezione di Pisa, 56127, Pisa, Italy
12
INFN-Sezione di Roma Tre, Rome, Italy
13
INFN-Sezione di Torino, Turin, Italy
14
INFN-Sezione di Trieste, Trieste, Italy
15
Institut für Experimentelle Teilchenphysik, Karlsruher Institut für Technologie, Karlsruhe, Germany
16
Institute of High Energy Physics, 1050, Vienna, Austria
17
Jožef Stefan Institute, Ljubljana, Slovenia
18
Ludwig Maximilians University, Munich, Germany
19
Max-Planck-Institut für Physik, Munich, Germany
20
Tel Aviv University, Tel Aviv, Israel
21
University of Bonn, Bonn, Germany
22
University of Melbourne, Melbourne, Australia
Received:
5
October
2021
Accepted:
27
January
2022
Published online:
8
February
2022
We present the novel implementation of a non-differentiable metric approximation and a corresponding loss-scheduling aimed at the search for new particles of unknown mass in high energy physics experiments. We call the loss-scheduling, based on the minimisation of a figure-of-merit related function typical of particle physics, a Punzi-loss function, and the neural network that utilises this loss function a Punzi-net. We show that the Punzi-net outperforms standard multivariate analysis techniques and generalises well to mass hypotheses for which it was not trained. This is achieved by training a single classifier that provides a coherent and optimal classification of all signal hypotheses over the whole search space. Our result constitutes a complementary approach to fully differentiable analyses in particle physics. We implemented this work using PyTorch and provide users full access to a public repository containing all the codes and a training example.
© The Author(s) 2022
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.
Funded by SCOAP3
