Towards a method to anticipate dark matter signals with deep learning at the LHC
We study several simplified dark matter (DM) models and their signatures at the LHC using neural networks. We focus on the usual monojet plus missing transverse energy channel, but to train the algorithms we organize the data in 2D histograms instead of event-by-event arrays. This results in a la...
| Main Author: | |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
SciPost
2022-02-01
|
| Series: | SciPost Physics |
| Online Access: | https://scipost.org/SciPostPhys.12.2.063 |
| _version_ | 1828853776681795584 |
|---|---|
| author | Ernesto Arganda, Anibal D. Medina, Andres D. Perez, Alejandro Szynkman |
| author_facet | Ernesto Arganda, Anibal D. Medina, Andres D. Perez, Alejandro Szynkman |
| author_sort | Ernesto Arganda, Anibal D. Medina, Andres D. Perez, Alejandro Szynkman |
| collection | DOAJ |
| description | We study several simplified dark matter (DM) models and their signatures at
the LHC using neural networks. We focus on the usual monojet plus missing
transverse energy channel, but to train the algorithms we organize the data in
2D histograms instead of event-by-event arrays. This results in a large
performance boost to distinguish between standard model (SM) only and SM plus
new physics signals. We use the kinematic monojet features as input data which
allow us to describe families of models with a single data sample. We found
that the neural network performance does not depend on the simulated number of
background events if they are presented as a function of $S/\sqrt{B}$, where
$S$ and $B$ are the number of signal and background events per histogram,
respectively. This provides flexibility to the method, since testing a
particular model in that case only requires knowing the new physics monojet
cross section. Furthermore, we also discuss the network performance under
incorrect assumptions about the true DM nature. Finally, we propose multimodel
classifiers to search and identify new signals in a more general way, for the
next LHC run. |
| first_indexed | 2024-12-13T00:20:06Z |
| format | Article |
| id | doaj.art-db19688a8eb148fc8df2ab89360c3d39 |
| institution | Directory Open Access Journal |
| issn | 2542-4653 |
| language | English |
| last_indexed | 2024-12-13T00:20:06Z |
| publishDate | 2022-02-01 |
| publisher | SciPost |
| record_format | Article |
| series | SciPost Physics |
| spelling | doaj.art-db19688a8eb148fc8df2ab89360c3d392022-12-22T00:05:38ZengSciPostSciPost Physics2542-46532022-02-0112206310.21468/SciPostPhys.12.2.063Towards a method to anticipate dark matter signals with deep learning at the LHCErnesto Arganda, Anibal D. Medina, Andres D. Perez, Alejandro SzynkmanWe study several simplified dark matter (DM) models and their signatures at the LHC using neural networks. We focus on the usual monojet plus missing transverse energy channel, but to train the algorithms we organize the data in 2D histograms instead of event-by-event arrays. This results in a large performance boost to distinguish between standard model (SM) only and SM plus new physics signals. We use the kinematic monojet features as input data which allow us to describe families of models with a single data sample. We found that the neural network performance does not depend on the simulated number of background events if they are presented as a function of $S/\sqrt{B}$, where $S$ and $B$ are the number of signal and background events per histogram, respectively. This provides flexibility to the method, since testing a particular model in that case only requires knowing the new physics monojet cross section. Furthermore, we also discuss the network performance under incorrect assumptions about the true DM nature. Finally, we propose multimodel classifiers to search and identify new signals in a more general way, for the next LHC run.https://scipost.org/SciPostPhys.12.2.063 |
| spellingShingle | Ernesto Arganda, Anibal D. Medina, Andres D. Perez, Alejandro Szynkman Towards a method to anticipate dark matter signals with deep learning at the LHC SciPost Physics |
| title | Towards a method to anticipate dark matter signals with deep learning at the LHC |
| title_full | Towards a method to anticipate dark matter signals with deep learning at the LHC |
| title_fullStr | Towards a method to anticipate dark matter signals with deep learning at the LHC |
| title_full_unstemmed | Towards a method to anticipate dark matter signals with deep learning at the LHC |
| title_short | Towards a method to anticipate dark matter signals with deep learning at the LHC |
| title_sort | towards a method to anticipate dark matter signals with deep learning at the lhc |
| url | https://scipost.org/SciPostPhys.12.2.063 |
| work_keys_str_mv | AT ernestoargandaanibaldmedinaandresdperezalejandroszynkman towardsamethodtoanticipatedarkmattersignalswithdeeplearningatthelhc |