Toward using GANs in astrophysical Monte-Carlo simulations
Accurate modelling of spectra produced by X-ray sources requires the use of Monte-Carlo simulations. These simulations need to evaluate physical processes, such as those occurring in accretion processes around compact objects by sampling a number of different probability distributions. This is compu...
Main Authors: | , , |
---|---|
Format: | Internet publication |
Language: | English |
Published: |
2024
|
_version_ | 1811139373462192128 |
---|---|
author | Isaac, A Armour, W Adámek, K |
author_facet | Isaac, A Armour, W Adámek, K |
author_sort | Isaac, A |
collection | OXFORD |
description | Accurate modelling of spectra produced by X-ray sources requires the use of Monte-Carlo simulations. These simulations need to evaluate physical processes, such as those occurring in accretion processes around compact objects by sampling a number of different probability distributions. This is computationally time-consuming and could be sped up if replaced by neural networks. We demonstrate, on an example of the Maxwell-Jüttner distribution that describes the speed of relativistic electrons, that the generative adversarial network (GAN) is capable of statistically replicating the distribution. The average value of the Kolmogorov-Smirnov test is 0.5 for samples generated by the neural network, showing that the generated distribution cannot be distinguished from the true distribution. |
first_indexed | 2024-09-25T04:05:03Z |
format | Internet publication |
id | oxford-uuid:534b096c-5ac3-41af-93df-0228ccffe99b |
institution | University of Oxford |
language | English |
last_indexed | 2024-09-25T04:05:03Z |
publishDate | 2024 |
record_format | dspace |
spelling | oxford-uuid:534b096c-5ac3-41af-93df-0228ccffe99b2024-05-20T16:04:35ZToward using GANs in astrophysical Monte-Carlo simulationsInternet publicationhttp://purl.org/coar/resource_type/c_7ad9uuid:534b096c-5ac3-41af-93df-0228ccffe99bEnglishSymplectic Elements2024Isaac, AArmour, WAdámek, KAccurate modelling of spectra produced by X-ray sources requires the use of Monte-Carlo simulations. These simulations need to evaluate physical processes, such as those occurring in accretion processes around compact objects by sampling a number of different probability distributions. This is computationally time-consuming and could be sped up if replaced by neural networks. We demonstrate, on an example of the Maxwell-Jüttner distribution that describes the speed of relativistic electrons, that the generative adversarial network (GAN) is capable of statistically replicating the distribution. The average value of the Kolmogorov-Smirnov test is 0.5 for samples generated by the neural network, showing that the generated distribution cannot be distinguished from the true distribution. |
spellingShingle | Isaac, A Armour, W Adámek, K Toward using GANs in astrophysical Monte-Carlo simulations |
title | Toward using GANs in astrophysical Monte-Carlo simulations |
title_full | Toward using GANs in astrophysical Monte-Carlo simulations |
title_fullStr | Toward using GANs in astrophysical Monte-Carlo simulations |
title_full_unstemmed | Toward using GANs in astrophysical Monte-Carlo simulations |
title_short | Toward using GANs in astrophysical Monte-Carlo simulations |
title_sort | toward using gans in astrophysical monte carlo simulations |
work_keys_str_mv | AT isaaca towardusinggansinastrophysicalmontecarlosimulations AT armourw towardusinggansinastrophysicalmontecarlosimulations AT adamekk towardusinggansinastrophysicalmontecarlosimulations |