簡易檢索 / 詳目顯示

回結果列表
研究生： 王純一
Wang, Chun-Yi
論文名稱： 圖神經網路應用於大軌跡半徑粒子的追蹤
Graph Neural Network for Large Radius Tracking
指導教授： 徐士傑
Hsu, Shih-Chieh
徐百嫻
Hsu, Pai-Hsien
口試委員: 張敬民
Cheung, King-Man
陳凱風
Chen, Kai-Feng
學位類別： 碩士
Master
系所名稱： 理學院 - 物理學系
Department of Physics
論文出版年： 2022
畢業學年度： 110
語文別： 英文
論文頁數： 34
中文關鍵詞： 機器學習 、圖神經網路 、長生命期粒子 、帶電粒子追蹤 、大軌跡半徑粒子追蹤 、超越標準模型
外文關鍵詞： Machine Learning, Graph Neural Network, Long-lived Particles, Charged Particle Tracking, Large Radius Tracking, Beyond the Standard Model
相關次數： 點閱：3  下載：0 
分享至:
查詢本校圖書館目錄 查詢臺灣博碩士論文知識加值系統 勘誤回報

    • 許多超越標準模型理論預測了長生命期粒子。這些粒子具有較長的半衰期，會移動一定距離後才產生衰變，其衰變的產物造成了大半徑軌跡。但此類大半徑軌跡若利用傳統建立於卡曼濾波的演算法進行重建，將需要大量額外的調整以及計算時間。近期相關研究表明，圖神經網路可以利用於從交互作用點生成的粒子軌跡重建，並展現了優異的效果以及計算時間。在本工作中，我們利用 ACTS 於通用粒子追蹤器模擬長生命期粒子事件，並使用圖神經網路進行重建以及分析其效能。作為結果，我們成功重建一般以及大軌跡半徑，且相比於一般軌跡，大軌跡半徑的重建並沒有顯著的效率下降。

    Long-lived particles are predicted by many Beyond Standard Model theories. They have longer lifetime, travel a distance before decaying to other particles, resulting in large radius tracks. However, such type of tracks are not easy to reconstruct with traditional algorithms based on the Kalman filter, which require large computational cost and dedicated tuning for finding such type of tracks. Recent studies show that Machine Learning-based particle track finding algorithm using graph neural network (GNN) achieves competitive physics and computing performance for tracks originated from collision point. In this work, we generate displaced track dataset under ACTS framework with generic detector and apply GNN-based algorithm on such dataset to study the performance of algorithm in this type of reconstruction problem. As the result, we reconstruct prompt and displaced tracks simultaneously with high track efficiency and no significant drop for highly displaced tracks.

    Abstract...................................................................ii Acknowledgements..........................................................iii 1 Introduction..............................................................1 1.1 ParticleTrackReconstruction.............................................1 1.2 Long-lived Particles and Beyond Standard Model Theories.................2 1.3 MachineLearning.........................................................3 1.4 GraphNeuralNetwork......................................................4 2 Simulated Data............................................................7 2.1 HeavyNeutralLeptonProcess...............................................7 2.2 DetectorandEventSimulation..............................................9 3 Exa.TrkX Pipeline........................................................11 3.1 GenericGraphConstruction...............................................12 3.2 EdgeFiltering..........................................................12 3.3 GraphNeuralNetworkEdgeClassification ..................................13 3.4 TrackReconstruction....................................................14 4 Model Training and Hyperparameter Tuning.................................16 4.1 TrainingProcedure .....................................................16 4.1.1 GraphConstruction....................................................17 4.1.2 EdgeFiltering........................................................17 4.1.3 GraphNeuralNetworkClassifier.........................................17 4.2 Hyperparameters........................................................17 4.2.1 GraphConstruction....................................................18 4.2.2 Filtering............................................................19 4.2.3 GraphNeuralNetworkClassifier ........................................19 4.2.4 TrackReconstruction..................................................19 5 Result...................................................................20 5.1 GraphConstructionandEdgeFiltering......................................20 5.2 GraphNeuralNetwork.....................................................20 5.3 TrackReconstruction....................................................21 6 Conclusion...............................................................25 Appendix...................................................................26 A Model Summary............................................................26 A.1 Embedding..............................................................26 A.2 Filter.................................................................27 A.3 GNNClassifier..........................................................28 B Online Resources.........................................................30 C Future Studies...........................................................31 References.................................................................32

    [1] “Trackml particle tracking challenge in kaggle.” https://www.kaggle.com/c/trackml-particle-identification, Jan. 2018.
    [2] S. Farrell, P. Calafiura, M. Mudigonda, Prabhat, D. Anderson, J.-R. Vlimant, S. Zheng, J. Bendavid, M. Spiropulu, G. Cerati, L. Gray, J. Kowalkowski, P. Spentzouris, and A. Tsaris, “Novel deep learning methods for track reconstruction,” 2018.
    [3] A. Strandlie and R. Frühwirth, “Track and vertex reconstruction: From classical to adaptive methods,” Rev. Mod. Phys., vol. 82, pp. 1419–1458, May 2010.
    [4] A. Collaboration, “Performance of the ATLAS track reconstruction algorithms in dense environments in LHC run 2,” The European Physical Journal C, vol. 77, oct 2017.
    [5] C. Collaboration, “Description and performance of track and primary-vertex reconstruction with the CMS tracker,” Journal of Instrumentation, vol. 9, pp. P10009–P10009, oct 2014.
    [6] T. Yanagida, “Horizontal Symmetry and Masses of Neutrinos,” Progress of Theoretical Physics, vol. 64, pp. 1103–1105, 09 1980.
    [7] G. R. Farrar and P. Fayet, “Phenomenology of the Production, Decay, and Detection of New Hadronic States Associated with Supersymmetry,” Phys. Lett. B, vol. 76, pp. 575–579, 1978.
    [8] T.Appelquist,H.-C.Cheng,andB.A.Dobrescu,“Boundsonuniversalextradimensions,” Phys. Rev. D, vol. 64, p. 035002, Jun 2001.
    [9] J. C. Pati and A. Salam, “Lepton number as the fourth ”color”,” Phys. Rev. D, vol. 10, pp. 275–289, Jul 1974.
    [10] “Performance of the reconstruction of large impact parameter tracks in the ATLAS inner detector,” tech. rep., CERN, Geneva, Jul 2017. All figures including auxiliary figures are available at https://atlas.web.cern.ch/Atlas/GROUPS/PHYSICS/PUBNOTES/ATL-PHYS-PUB-2017-014.
    [11] D. E. Rumelhart, G. E. Hinton, and R. J. Williams, “Learning representations by backpropagating errors,” Nature, vol. 323, pp. 533–536, Oct 1986.
    [12] J. Zhou, G. Cui, S. Hu, Z. Zhang, C. Yang, Z. Liu, L. Wang, C. Li, and M. Sun, “Graph neural networks: A review of methods and applications,” 2018.
    [13] P. Veličković, G. Cucurull, A. Casanova, A. Romero, P. Liò, and Y. Bengio, “Graph attention networks,” 2017. 32
    [14] S. Farrell et al., “The HEP.TrkX Project: deep neural networks for HL-LHC online and offline tracking,” in Proceedings, Connecting The Dots / Intelligent Tracker (CTD/WIT 2017): Orsay, France, March 6-9, 2017, vol. 150, p. 00003, 2017.
    [15] X. Ju et al., “Graph Neural Networks for Particle Reconstruction in High Energy Physics detectors,” in 33rd Annual Conference on Neural Information Processing Systems, 3 2020.
    [16] N. Choma, D. Murnane, X. Ju, P. Calafiura, S. Conlon, S. Farrell, Prabhat, G. Cerati, L. Gray, T. Klijnsma, J. Kowalkowski, P. Spentzouris, J.-R. Vlimant, M. Spiropulu, A. Aurisano, J. Hewes, A. Tsaris, K. Terao, and T. Usher, “Track seeding and labelling with embedded-space graph neural networks,” 2020.
    [17] X. Ju, D. Murnane, P. Calafiura, N. Choma, S. Conlon, S. Farrell, Y. Xu, M. Spiropulu, J.-R. Vlimant, A. Aurisano, J. Hewes, G. Cerati, L. Gray, T. Klijnsma, J. Kowalkowski, M. Atkinson, M. Neubauer, G. DeZoort, S. Thais, A. Chauhan, A. Schuy, S.-C. Hsu, A. Ballow, and A. Lazar, “Performance of a geometric deep learning pipeline for hl-lhc particle tracking,” The European Physical Journal C, vol. 81, Oct 2021.
    [18] C. Biscarat, S. Caillou, C. Rougier, J. Stark, and J. Zahreddine, “Towards a realistic track reconstruction algorithm based on graph neural networks for the hl-lhc,” 2021.
    [19] D. A. Trischuk, “Displaced Vertex Search For Heavy Neutral Leptons Using The ATLAS Detector,” 2022. Presented 19 Jan 2022.
    [20] M. Drewes, B. Garbrecht, D. Gueter, and J. Klarić, “Testing the low scale seesaw and leptogenesis,” Journal of High Energy Physics, vol. 2017, aug 2017.
    [21] T. Asaka, S. Blanchet, and M. Shaposhnikov, “The νMSM, dark matter and neutrino masses,” Physics Letters B, vol. 631, pp. 151–156, dec 2005.
    [22] T.Sjöstrand,S.Ask,J.R.Christiansen,R.Corke,N.Desai,P.Ilten,S.Mrenna,S.Prestel, C. O. Rasmussen, and P. Z. Skands, “An introduction to PYTHIA 8.2,” Computer Physics Communications, vol. 191, pp. 159–177, jun 2015.
    [23] X. Ai, “Acts: A common tracking software,” 2019.
    [24] K. Edmonds, S. Fleischmann, T. Lenz, C. Magass, J. Mechnich, and A. Salzburger, “The Fast ATLAS Track Simulation (FATRAS),” tech. rep., CERN, Geneva, Mar 2008. All figures including auxiliary figures are available at https://atlas.web.cern.ch/Atlas/GROUPS/PHYSICS/PUBNOTES/ATL-SOFT-PUB- 2008-001.
    [25] Exa.TrkX, “ HEP advanced tracking algorithms at the exascale,” 2019.
    [26] Z. Wu, S. Pan, F. Chen, G. Long, C. Zhang, and P. S. Yu, “A comprehensive survey on graph neural networks,” IEEE Transactions on Neural Networks and Learning Systems, vol. 32, p. 4–24, Jan 2021.
    [27] M. Ester, H.-P. Kriegel, J. Sander, and X. Xu, “A density-based algorithm for discovering clusters in large spatial databases with noise,” in Kdd, pp. 226–231, AAAI Press, 1996.
    [28] I. Loshchilov and F. Hutter, “Fixing weight decay regularization in adam,” CoRR, vol. abs/1711.05101, 2017. 33
    [29] S. J. Reddi, S. Kale, and S. Kumar, “On the convergence of adam and beyond,” 2019.
    [30] A. Heintz et al., “Accelerated Charged Particle Tracking with Graph Neural Networks on FPGAs,” in 34th Conference on Neural Information Processing Systems, 11 2020.
    [31] A. Elabd et al., “Graph Neural Networks for Charged Particle Tracking on FPGAs,” 12 2021.
    [32] K. He, X. Zhang, S. Ren, and J. Sun, “Deep residual learning for image recognition,” in 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), pp. 770– 778, 2016.
    [33] C.Allaire, P.Gessinger, J.Hdrinka, M.Kiehn, F.Kimpel, J.Niermann, A.Salzburger, and S. Sevova, “Opendatadetector,” Apr. 2022.

    QR CODE