Separation of π0/γ in an Electromagnetic Sampling Calorimeter with Deep Learning Structures

Year 2021, Issue: 32, 1175 - 1180, 31.12.2021

Güral Aydın Hasan Sarıgül

https://doi.org/10.31590/ejosat.1041107

Abstract

In this study, it was investigated how effective the use of deep learning artificial neural networks can be in distinguishing neutral pion and single photon in an electromagnetic sampling calorimeter. The sampling calorimeter was constructed in the form of a 9×9 matrix array with the Geant4 simulation program. Identification of the particles was carried out by using the shower images created by the neutral pion and a single photon at different energies in the calorimeter. First, differences in shower images were observed using image parameters. Then, the topologies created by the shower images were used as input parameters in deep learning structures to distinguish the particles. It has been observed that very high signal efficiency and background rejection values can be achieved under the specified simulation conditions with machine learning.

Keywords

Artificial Neural Network, Electromagnetic Calorimeter, Photon Interaction, Particle Identification

Derin Öğrenme Yapıları ile Elektromanyetik Örnekleme Kalorimetresinde π0/γ Ayırt Edilmesi

Abstract

Bu çalışmada, bir elektromanyetik örnekleme kalorimetresinde nötr pion ve tek fotonun ayırt edilmesinde derin öğrenme yapay sinir ağlarının kullanılmasının ne kadar etkili olabileceği araştırılmıştır. Örnekleme kalorimetresi 9×9 matris dizini şeklinde Geant4 benzetim programı ile oluşturulmuştur. Nötr pion ve tek fotonun farklı enerjilerde kalorimetrede oluşturduğu duş görüntüleri kullanılarak parçacıkların tanımlanması gerçekleştirilmiştir. İlk olarak, görüntü parametreleri kullanılarak duş görüntülerindeki farklılıklar gözlemlenmiştir. Daha sonra, derin öğrenme yapıları içerisinde duş görüntülerinin oluşturduğu topolojiler giriş parametreleri olarak kullanılarak parçacıkların ayırt edilmesine çalışılmıştır. Makine öğrenmesi ile birlikte belirtilen benzetim koşullarında oldukça yüksek seviyede sinyal verimliliği ve arkaplan reddi değerlerine ulaşılabileceği görülmüştür.

Keywords

Yapay Sinir Ağları, Elektromanyetik Kalorimetresi, Foton Etkileşmesi, Parçacık Tanımlanması

References

• Agostinelli, S., Allison, J., Amako, K., Apostolakis, J., Araujo, H., Arce, P. … & Zschiesche, D. (2003). Geant4—a simulation toolkit. Nuclear Instruments and Methods in Physics Research Section A, 506 (3), 250-303.
• Allison J., Amako K., Apostolakis J., Araujo H., Arce P., Asai M. … & Yoshida, H. (2006). Geant4 developments and applications. IEEE Transactions on Nuclear Science, 53 (1), 270-278.
• Allison J., Amako K., Apostolakis J., Arce P., Asai M., Aso T. … & Yoshida, H. (2016). Recent developments in Geant4. Nuclear Instruments and Methods in Physics Research Section A, 835, 186-225.
• Aydın, G., Sarıgül. M., & Sarıgül. H. (2020). Position resolution study at high energies of a sampling electromagnetic calorimeter whose active material is a scintillator with Peroxide-cured polysiloxane base. Nuclear Instruments and Methods in Physics Research A, 955 (Mart), 163341.
• Deng, L., Li, J., Huang, J.T., Yao, K., Yu, D., Seide, F. … & Acero, A. (2013). Recent advances in deep learning for speech research at Microsoft. IEEE International Conference on Acoustics, Speech and Signal Processing, ICASSP, Vancouver: IEEE, 13859384.
• Diemoz, M. (CMS ECAL İşbirliği Adına). (2007). The electromagnetic calorimeter of the CMS experiment. Nuclear Instruments and Methods in Physics Research Section A, 581 (1-2), 380-383.
• Glorot, X., Bordes, A., & Bengio, Y. (2011). Deep Sparse Rectifier Neural Networks. Proceedings of the Fourteenth International Conference on Artificial Intelligence and Statistics, PMLR, Fort Lauderdale, 15, 315-323.
• Graves, A., Mohamed, A., & Hinton, G. (2013). Speech recognition with deep recurrent neural networks. International Conference on Acoustics, Speech and Signal Processing, ICASSP, Vancouver: IEEE, 6645–6649.
• Hinton, G.E., & Salakhutdinov, R.R. (2006). Reducing the dimensionality of data with neural network. Science, 313 (5786), 504–507.
• Hinton, G., Deng, L., Yu, D., Dahl, G.E., Mohamed, A., Jaitly, N. … & Kingsbury, B. (2012). Deep neural networks for acoustic modeling in speech recognition: The shared views of four research groups. Signal Processing Magazine, IEEE, 29 (6), 82–97.
• Keras. (2021). https://keras.io. (Erişim 15 Mart 2021).
• Kingma, D.P., & Ba, J. (2017). Adam: A Method for Stochastic Optimization. arXiv:1412.6980v9 [cs.LG].
• Krizhevsky, A., Sutskever, I., & Hinton, G.E. (2017). Imagenet classification with deep convolutional neural networks. Communications of the ACM, 60 (6), 84-90.
• Roy, A., Jain, S., Banerjee, S., Bhattacharya, S., & Majumfer, G. (2016). Simulation of π^0-γ separation study for proposed CMS forward electromagnetic calorimeter, J. Phys. Conf. Ser., 759, 012074.
• Roy, A., Jain, S., Banerjee, S., Bhattacharya, S., & Majumfer, G. (2017). Simulation study of energy resolution, position resolution and π0-γ separation of a sampling electromagnetic calorimeter at high energies. Journal of Instrumentation, 12, P07013.
• Tensorflow. (2021). https://www.tensorflow.org. (Erişim 15 Mart 2021).