Data driven background estimation in HEP using generative adversarial networks
Abstract Data-driven methods are widely used to overcome shortcomings of Monte Carlo simulations (lack of statistics, mismodeling of processes, etc.) in experimental high energy physics. A precise description of background processes is crucial to reach the optimal sensitivity for a measurement. Howe...
| Main Authors: | , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
SpringerOpen
2023-03-01
|
| Series: | European Physical Journal C: Particles and Fields |
| Online Access: | https://doi.org/10.1140/epjc/s10052-023-11347-8 |
| _version_ | 1797836321128972288 |
|---|---|
| author | Victor Lohezic Mehmet Ozgur Sahin Fabrice Couderc Julie Malcles |
| author_facet | Victor Lohezic Mehmet Ozgur Sahin Fabrice Couderc Julie Malcles |
| author_sort | Victor Lohezic |
| collection | DOAJ |
| description | Abstract Data-driven methods are widely used to overcome shortcomings of Monte Carlo simulations (lack of statistics, mismodeling of processes, etc.) in experimental high energy physics. A precise description of background processes is crucial to reach the optimal sensitivity for a measurement. However, the selection of the control region used to describe the background process in a region of interest biases the distribution of some physics observables, rendering the use of such observables impossible in a physics analysis. Rather than discarding these events and/or observables, we propose a novel method to generate physics objects compatible with the region of interest and properly describing the correlations with the rest of the event properties. We use a generative adversarial network (GAN) for this task, as GANs are among the best generator models for various applications. We illustrate the method by generating a new misidentified photon for the $$\gamma + \textrm{jets}$$ γ + jets background of the $$\textrm{H}\rightarrow \gamma \gamma $$ H → γ γ analysis at the CERN LHC, and demonstrate that this GAN generator is able to produce a coherent object correlated with the different properties of the rest of the event. |
| first_indexed | 2024-04-09T15:08:01Z |
| format | Article |
| id | doaj.art-f2351dae3c2d466cb1a7ddac02807481 |
| institution | Directory Open Access Journal |
| issn | 1434-6052 |
| language | English |
| last_indexed | 2024-04-09T15:08:01Z |
| publishDate | 2023-03-01 |
| publisher | SpringerOpen |
| record_format | Article |
| series | European Physical Journal C: Particles and Fields |
| spelling | doaj.art-f2351dae3c2d466cb1a7ddac028074812023-04-30T11:25:48ZengSpringerOpenEuropean Physical Journal C: Particles and Fields1434-60522023-03-018331910.1140/epjc/s10052-023-11347-8Data driven background estimation in HEP using generative adversarial networksVictor Lohezic0Mehmet Ozgur Sahin1Fabrice Couderc2Julie Malcles3IRFU, CEA, Université Paris-SaclayIRFU, CEA, Université Paris-SaclayIRFU, CEA, Université Paris-SaclayIRFU, CEA, Université Paris-SaclayAbstract Data-driven methods are widely used to overcome shortcomings of Monte Carlo simulations (lack of statistics, mismodeling of processes, etc.) in experimental high energy physics. A precise description of background processes is crucial to reach the optimal sensitivity for a measurement. However, the selection of the control region used to describe the background process in a region of interest biases the distribution of some physics observables, rendering the use of such observables impossible in a physics analysis. Rather than discarding these events and/or observables, we propose a novel method to generate physics objects compatible with the region of interest and properly describing the correlations with the rest of the event properties. We use a generative adversarial network (GAN) for this task, as GANs are among the best generator models for various applications. We illustrate the method by generating a new misidentified photon for the $$\gamma + \textrm{jets}$$ γ + jets background of the $$\textrm{H}\rightarrow \gamma \gamma $$ H → γ γ analysis at the CERN LHC, and demonstrate that this GAN generator is able to produce a coherent object correlated with the different properties of the rest of the event.https://doi.org/10.1140/epjc/s10052-023-11347-8 |
| spellingShingle | Victor Lohezic Mehmet Ozgur Sahin Fabrice Couderc Julie Malcles Data driven background estimation in HEP using generative adversarial networks European Physical Journal C: Particles and Fields |
| title | Data driven background estimation in HEP using generative adversarial networks |
| title_full | Data driven background estimation in HEP using generative adversarial networks |
| title_fullStr | Data driven background estimation in HEP using generative adversarial networks |
| title_full_unstemmed | Data driven background estimation in HEP using generative adversarial networks |
| title_short | Data driven background estimation in HEP using generative adversarial networks |
| title_sort | data driven background estimation in hep using generative adversarial networks |
| url | https://doi.org/10.1140/epjc/s10052-023-11347-8 |
| work_keys_str_mv | AT victorlohezic datadrivenbackgroundestimationinhepusinggenerativeadversarialnetworks AT mehmetozgursahin datadrivenbackgroundestimationinhepusinggenerativeadversarialnetworks AT fabricecouderc datadrivenbackgroundestimationinhepusinggenerativeadversarialnetworks AT juliemalcles datadrivenbackgroundestimationinhepusinggenerativeadversarialnetworks |