Physics-Informed Deep Learning for Three Dimensional Black Holes
In this paper, we have designed an artificial neural network architecture to produce metric field of planar BTZ and quintessence black holes applying a data-driven approach andleveraging holography principle (according to AdS/DL (Anti de Sitter/ Deep Learning) correspondence given by [1]). Data has...
Saved in:
| Main Authors: | , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Damghan university
2023-12-01
|
| Series: | Iranian Journal of Astronomy and Astrophysics |
| Subjects: | |
| Online Access: | https://ijaa.du.ac.ir/article_387_f3bef28fe0a5faab9271b484e81e4235.pdf |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1841545164024184832 |
|---|---|
| author | Emad Yaraie Hossein Ghaffarnejad Mohammad Farsam |
| author_facet | Emad Yaraie Hossein Ghaffarnejad Mohammad Farsam |
| author_sort | Emad Yaraie |
| collection | DOAJ |
| description | In this paper, we have designed an artificial neural network architecture to produce metric field of planar BTZ and quintessence black holes applying a data-driven approach andleveraging holography principle (according to AdS/DL (Anti de Sitter/ Deep Learning) correspondence given by [1]). Data has been collected by choosing minimally coupled massive scalar field with quantum fluctuations and we try to process two emergent and ground-truth metrics versus the holographic parameter which plays the role of depth of the neural network. Loss or error function which shows rate of deviation of these two metrics in presence of penalty regularization term reaches to its minimum value when values of the learning rate approach to the observed steepest gradient point. Values of the regularization or penalty term of the quantum scalar field has critical role to matching this two mentioned metric. Also, we design an algorithm which helps us to find optimum value for learning parameter and finally, we understand that loss function convergence heavily depends on the number of epochs and learning rate. |
| format | Article |
| id | doaj-art-e1fc435fb91a4f33bdf5153f6df7b77c |
| institution | Kabale University |
| issn | 2322-4924 2383-403X |
| language | English |
| publishDate | 2023-12-01 |
| publisher | Damghan university |
| record_format | Article |
| series | Iranian Journal of Astronomy and Astrophysics |
| spelling | doaj-art-e1fc435fb91a4f33bdf5153f6df7b77c2025-01-12T10:03:51ZengDamghan universityIranian Journal of Astronomy and Astrophysics2322-49242383-403X2023-12-0110433535610.22128/ijaa.2023.694.1150387Physics-Informed Deep Learning for Three Dimensional Black HolesEmad Yaraie0Hossein Ghaffarnejad1Mohammad Farsam2Instituut-Lorentz for Theoretical Physics, ITP, Leiden UniversityFaculty of Physics, Semnan Universiy, Semnan, Iran, 35131-19111Instituut-Lorentz for Theoretical Physics, ITP, Leiden UniversityIn this paper, we have designed an artificial neural network architecture to produce metric field of planar BTZ and quintessence black holes applying a data-driven approach andleveraging holography principle (according to AdS/DL (Anti de Sitter/ Deep Learning) correspondence given by [1]). Data has been collected by choosing minimally coupled massive scalar field with quantum fluctuations and we try to process two emergent and ground-truth metrics versus the holographic parameter which plays the role of depth of the neural network. Loss or error function which shows rate of deviation of these two metrics in presence of penalty regularization term reaches to its minimum value when values of the learning rate approach to the observed steepest gradient point. Values of the regularization or penalty term of the quantum scalar field has critical role to matching this two mentioned metric. Also, we design an algorithm which helps us to find optimum value for learning parameter and finally, we understand that loss function convergence heavily depends on the number of epochs and learning rate.https://ijaa.du.ac.ir/article_387_f3bef28fe0a5faab9271b484e81e4235.pdfmachine learning, deep learning, black holes, three dimensions, btzoptimization, loss function |
| spellingShingle | Emad Yaraie Hossein Ghaffarnejad Mohammad Farsam Physics-Informed Deep Learning for Three Dimensional Black Holes Iranian Journal of Astronomy and Astrophysics machine learning, deep learning, black holes, three dimensions, btz optimization, loss function |
| title | Physics-Informed Deep Learning for Three Dimensional Black Holes |
| title_full | Physics-Informed Deep Learning for Three Dimensional Black Holes |
| title_fullStr | Physics-Informed Deep Learning for Three Dimensional Black Holes |
| title_full_unstemmed | Physics-Informed Deep Learning for Three Dimensional Black Holes |
| title_short | Physics-Informed Deep Learning for Three Dimensional Black Holes |
| title_sort | physics informed deep learning for three dimensional black holes |
| topic | machine learning, deep learning, black holes, three dimensions, btz optimization, loss function |
| url | https://ijaa.du.ac.ir/article_387_f3bef28fe0a5faab9271b484e81e4235.pdf |
| work_keys_str_mv | AT emadyaraie physicsinformeddeeplearningforthreedimensionalblackholes AT hosseinghaffarnejad physicsinformeddeeplearningforthreedimensionalblackholes AT mohammadfarsam physicsinformeddeeplearningforthreedimensionalblackholes |