- Physical Review Journals
All JournalsPhysics Magazine

Physical Review D

Reuse & Permissions

It is not necessary to obtain permission to reuse this article or its components as it is available under the terms of the Creative Commons Attribution 4.0 International license. This license permits unrestricted use, distribution, and reproduction in any medium, provided attribution to the author(s) and the published article's title, journal citation, and DOI are maintained. Please note that some figures may have been included with permission from other third parties. It is your responsibility to obtain the proper permission from the rights holder directly for these figures.

Export citation

Export citation

Choose format for download:

Download Citation
  • Letter
  • Open Access

Machine learning amplitudes for faster event generation

Fady Bishara1,* and Marc Montull1,2,3,†

  • 1Deutsches Elektronen-Synchrotron DESY, Notkestraße 85, 22607 Hamburg, Germany
  • 2Paul Scherrer Institut, Forschungsstraße 111, 5232 Villigen PSI, Switzerland
  • 3Physik-Institut, Universität Zürich, Winterthurerstrasse 190, CH-8057 Zürich, Switzerland

  • *fady.bishara@desy.de
  • †marc.montull@gmail.com
PDF

Phys. Rev. D 107, L071901 – Published 11 April, 2023

DOI: https://doi.org/10.1103/PhysRevD.107.L071901

Abstract

We propose to replace the exact amplitudes used in Monte Carlo event generators for trained machine learning regressors, with the aim of speeding up the evaluation of slow amplitudes. As a proof of concept, we study the process gg→ZZ, whose leading-order amplitude is loop induced. We show that gradient boosting machines like xgboost can predict the fully differential distributions with errors below 0.1%, and with prediction times O(103) faster than the evaluation of the exact function. This is achieved with training times ∼ 23  minutes and regressors of size ≲ 22  Mb. We also find that xgboost performs well over the entire phase space, while interpolation gives much larger errors in regions where the function is peaked. These results suggest a possible new avenue to speed up Monte Carlo event generators.

View figure in article

Physics Subject Headings (PhySH)

Article Text

Supplemental Material

References (90)

  1. J. Alwall, R. Frederix, S. Frixione, V. Hirschi, F. Maltoni, O. Mattelaer, H. S. Shao, T. Stelzer, P. Torrielli, and M. Zaro, The automated computation of tree-level and next-to-leading order differential cross sections, and their matching to parton shower simulations, J. High Energy Phys. 07 (2014) 079.
  2. V. Hirschi and O. Mattelaer, Automated event generation for loop-induced processes, J. High Energy Phys. 10 (2015) 146.
  3. R. Boughezal, J. M. Campbell, R. K. Ellis, C. Focke, W. Giele, X. Liu, F. Petriello, and C. Williams, Color singlet production at NNLO in MCFM, Eur. Phys. J. C 77, 7 (2017).
  4. J. Campbell and T. Neumann, Precision phenomenology with MCFM, J. High Energy Phys. 12 (2019) 034.
  5. Z. Nagy, Next-to-leading order calculation of three jet observables in hadron hadron collision, Phys. Rev. D 68, 094002 (2003).
  6. Z. Nagy, Three-Jet Cross Sections in Hadron-Hadron Collisions at Next-To-Leading Order, Phys. Rev. Lett. 88, 122003 (2002).
  7. R. Gavin, Y. Li, F. Petriello, and S. Quackenbush, FEWZ 2.0: A code for hadronic Z production at next-to-next-to-leading order, Comput. Phys. Commun. 182, 2388 (2011).
  8. S. Camarda et al., DYTurbo: Fast predictions for Drell–Yan processes, Eur. Phys. J. C 80, 251 (2020).
  9. S. Catani, L. Cieri, G. Ferrera, D. de Florian, and M. Grazzini, Vector boson production at hadron colliders: A fully exclusive QCD calculation at NNLO, Phys. Rev. Lett. 103, 082001 (2009).
  10. S. Catani, D. de Florian, G. Ferrera, and M. Grazzini, Vector boson production at hadron colliders: Transverse-momentum resummation and leptonic decay, J. High Energy Phys. 12 (2015) 047.
  11. D. de Florian, G. Ferrera, M. Grazzini, and D. Tommasini, Transverse-momentum resummation: Higgs boson production at the Tevatron and the LHC, J. High Energy Phys. 11 (2011) 064.
  12. S. Actis, A. Denner, L. Hofer, J.-N. Lang, A. Scharf, and S. Uccirati, RECOLA: REcursive Computation of One-Loop Amplitudes, Comput. Phys. Commun. 214, 140 (2017).
  13. V. Hirschi, R. Frederix, S. Frixione, M. V. Garzelli, F. Maltoni, and R. Pittau, Automation of one-loop QCD corrections, J. High Energy Phys. 05 (2011) 044.
  14. Z. Li, J. Wang, Q.-S. Yan, and X. Zhao, Efficient numerical evaluation of Feynman integrals, Chin. Phys. C 40, 033103 (2016).
  15. L. Naterop, A. Signer, and Y. Ulrich, handyG—Rapid numerical evaluation of generalised polylogarithms in Fortran, Comput. Phys. Commun. 253, 107165 (2020).
  16. T. Peraro, FiniteFlow: Multivariate functional reconstruction using finite fields and dataflow graphs, J. High Energy Phys. 07 (2019) 031.
  17. Z. Bern, L. J. Dixon, F. Febres Cordero, S. Höche, H. Ita, D. A. Kosower, and D. Maitre, Ntuples for NLO events at hadron colliders, Comput. Phys. Commun. 185, 1443 (2014).
  18. V. Bertone, S. Carrazza, and N. P. Hartland, APFELgrid: A high performance tool for parton density determinations, Comput. Phys. Commun. 212, 205 (2017).
  19. D. Maitre, G. Heinrich, and M. Johnson, N(N)LO event files: Applications and prospects, Proc. Sci. LL2016 (2016) 016.
  20. M. Spira, Higgs Boson production and decay at hadron colliders, Prog. Part. Nucl. Phys. 95, 98 (2017).
  21. S. Otten, K. Rolbiecki, S. Caron, J.-S. Kim, R. Ruiz De Austri, and J. Tattersall, DeepXS: Fast approximation of MSSM electroweak cross sections at NLO, Eur. Phys. J. C 80, 12 (2020).
  22. H. A. Chawdhry, M. L. Czakon, A. Mitov, and R. Poncelet, NNLO QCD corrections to three-photon production at the LHC, J. High Energy Phys. 02 (2020) 057.
  23. M. Bellagente, A. Butter, G. Kasieczka, T. Plehn, and R. Winterhalder, How to GAN away detector effects, SciPost Phys. 8, 070 (2020).
  24. A. Andreassen and B. Nachman, Neural networks for full phase-space reweighting and parameter tuning, Phys. Rev. D 101, 091901 (2020).
  25. B. Nachman, A guide for deploying Deep Learning in LHC searches: How to achieve optimality and account for uncertainty, SciPost Phys. 8, 090 (2020).
  26. J. Bendavid, Efficient Monte Carlo integration using boosted decision trees and generative deep neural networks, arXiv:1707.00028.
  27. S. Otten, S. Caron, W. de Swart, M. van Beekveld, L. Hendriks, C. van Leeuwen, D. Podareanu, R. Ruiz de Austri, and R. Verheyen, Event generation and statistical sampling for physics with deep generative models and a density information buffer, Nat. Commun. 12, 2985 (2021).
  28. M. D. Klimek and M. Perelstein, Neural network-based approach to phase space integration, SciPost Phys. 9, 053 (2020).
  29. E. Bothmann and L. Debbio, Reweighting a parton shower using a neural network: The final-state case, J. High Energy Phys. 01 (2019) 033.
  30. R. D. Ball et al. (NNPDF Collaboration), Parton distributions from high-precision collider data, Eur. Phys. J. C 77, 663 (2017).
  31. R. D. Ball et al. (NNPDF), Parton distributions for the LHC Run II, J. High Energy Phys. 04 (2015) 040.
  32. R. Santos, M. Nguyen, J. Webster, S. Ryu, J. Adelman, S. Chekanov, and J. Zhou, Machine learning techniques in searches for tt¯h in the h→bb¯ decay channel, J. Instrum. 12, P04014 (2017).
  33. M. Farina, Y. Nakai, and D. Shih, Searching for new physics with deep autoencoders, Phys. Rev. D 101, 075021 (2020).
  34. R. T. D’Agnolo and A. Wulzer, Learning new physics from a machine, Phys. Rev. D 99, 015014 (2019).
  35. A. Blance, M. Spannowsky, and P. Waite, Adversarially-trained autoencoders for robust unsupervised new physics searches, J. High Energy Phys. 10 (2019) 047.
  36. O. Cerri, T. Q. Nguyen, M. Pierini, M. Spiropulu, and J.-R. Vlimant, Variational autoencoders for new physics mining at the large hadron collider, J. High Energy Phys. 05 (2019) 036.
  37. S. Brochet, C. Delaere, B. François, V. Lemaître, A. Mertens, A. Saggio, M. Vidal Marono, and S. Wertz, MoMEMta, a modular toolkit for the Matrix Element Method at the LHC, Eur. Phys. J. C 79, 126 (2019).
  38. S. Banerjee, R. S. Gupta, J. Y. Reiness, S. Seth, and M. Spannowsky, Towards the ultimate differential SMEFT analysis, J. High Energy Phys. 09 (2020) 170.
  39. P. T. Komiske, E. M. Metodiev, and M. D. Schwartz, Deep learning in color: towards automated quark/gluon jet discrimination, J. High Energy Phys. 01 (2017) 110.
  40. A. Coccaro, M. Pierini, L. Silvestrini, and R. Torre, The DNNLikelihood: Enhancing likelihood distribution with Deep Learning, Eur. Phys. J. C 80, 664 (2020).
  41. A. Andreassen, I. Feige, C. Frye, and M. D. Schwartz, JUNIPR: A Framework for Unsupervised Machine Learning in Particle Physics, Eur. Phys. J. C 79, 102 (2019).
  42. A. Andreassen, I. Feige, C. Frye, and M. D. Schwartz, Binary JUNIPR: An interpretable probabilistic model for discrimination, Phys. Rev. Lett. 123, 182001 (2019).
  43. L. de Oliveira, M. Kagan, L. Mackey, B. Nachman, and A. Schwartzman, Jet-images—deep learning edition, J. High Energy Phys. 07 (2016) 069.
  44. E. M. Metodiev, B. Nachman, and J. Thaler, Classification without labels: Learning from mixed samples in high energy physics, J. High Energy Phys. 10 (2017) 174.
  45. J. H. Collins, K. Howe, and B. Nachman, Anomaly Detection for Resonant New Physics with Machine Learning, Phys. Rev. Lett. 121, 241803 (2018).
  46. J. H. Collins, K. Howe, and B. Nachman, Extending the search for new resonances with machine learning, Phys. Rev. D 99, 014038 (2019).
  47. A. Andreassen, P. T. Komiske, E. M. Metodiev, B. Nachman, and J. Thaler, OmniFold: A Method to Simultaneously Unfold All Observables, Phys. Rev. Lett. 124, 182001 (2020).
  48. M. Paganini, L. de Oliveira, and B. Nachman, Accelerating Science with Generative adversarial Networks: An Application to 3D Particle Showers in Multilayer Calorimeters, Phys. Rev. Lett. 120, 042003 (2018).
  49. M. Paganini, L. de Oliveira, and B. Nachman, CaloGAN : Simulating 3D high energy particle showers in multilayer electromagnetic calorimeters with generative adversarial networks, Phys. Rev. D 97, 014021 (2018).
  50. A. Chakraborty, S. H. Lim, and M. M. Nojiri, Interpretable deep learning for two-prong jet classification with jet spectra, J. High Energy Phys. 07 (2019) 135.
  51. S. H. Lim and M. M. Nojiri, Spectral analysis of jet substructure with neural networks: boosted higgs case, J. High Energy Phys. 10 (2018) 181.
  52. G. Luisoni, S. Poslavsky, and Y. Schroder, Track 3: Computations in theoretical physics—techniques and methods, J. Phys. Conf. Ser. 762, 012077 (2016).
  53. A. Buckley, Computational challenges for MC event generation, J. Phys. Conf. Ser. 1525, 012023 (2020).
  54. M. Abdughani, J. Ren, L. Wu, J. M. Yang, and J. Zhao, Supervised deep learning in high energy phenomenology: A mini review, Commun. Theor. Phys. 71, 955 (2019).
  55. K. Albertsson et al., Machine learning in high energy physics community white paper, J. Phys. Conf. Ser. 1085, 022008 (2018).
  56. J. Albrecht et al. (HEP Software Foundation Collaboration), A roadmap for HEP software and computing R&D for the 2020s, Comput. Softw. Big Sci. 3, 7 (2019).
  57. S. Carrazza, Machine learning challenges in theoretical HEP, J. Phys. Conf. Ser. 1085, 022003 (2018).
  58. A. J. Larkoski, I. Moult, and B. Nachman, Jet substructure at the Large Hadron Collider: A review of recent advances in theory and machine learning, Phys. Rep. 841, 1 (2020).
  59. D. Bourilkov, Machine and deep learning applications in particle physics, Int. J. Mod. Phys. A 34, 1930019 (2020).
  60. J. Brehmer, F. Kling, I. Espejo, and K. Cranmer, MadMiner: Machine learning-based inference for particle physics, Comput. Softw. Big Sci. 4, 3 (2020).
  61. J. Brehmer, S. Dawson, S. Homiller, F. Kling, and T. Plehn, Benchmarking simplified template cross sections in WH production, J. High Energy Phys. 11 (2019) 034.
  62. G. Carleo, I. Cirac, K. Cranmer, L. Daudet, M. Schuld, N. Tishby, L. Vogt-Maranto, and L. Zdeborová, Machine learning and the physical sciences, Rev. Mod. Phys. 91, 045002 (2019).

  63. A relevant example is diboson mediated four-lepton production, pp→V1V2→4ℓ. More specifically, Table 11 in the journal version of Ref. [64] shows that generating pp→4e at NNLO takes more than a thousand times longer than at NLO.

  64. M. Grazzini, S. Kallweit, and M. Wiesemann, Fully differential NNLO computations with MATRIX, Eur. Phys. J. C 78, 537 (2018).
  65. See Supplemental Material at http://link.aps.org/supplemental/10.1103/PhysRevD.107.L071901 for additional details regarding the shape of the squared matrix element, comparison with interpolation, and the continuity of the approximated function across the boundaries of the phase-space subregions.
  66. F. Cascioli, T. Gehrmann, M. Grazzini, S. Kallweit, P. Maierhöfer, A. von Manteuffel, S. Pozzorini, D. Rathlev, L. Tancredi, and E. Weihs, ZZ production at hadron colliders in NNLO QCD, Phys. Lett. B 735, 311 (2014).
  67. V. Khachatryan et al. (CMS Collaboration), Searches for invisible decays of the Higgs boson in pp collisions at s=7,8,and 13TeV, J. High Energy Phys. 02 (2017) 135.
  68. M. Aaboud et al. (ATLAS Collaboration), Observation of H→bb¯ decays and VH production with the ATLAS detector, Phys. Lett. B 786, 59 (2018).
  69. A. M. Sirunyan et al. (CMS Collaboration), Observation of Higgs boson Decay to Bottom Quarks, Phys. Rev. Lett. 121, 121801 (2018).
  70. S. Banerjee, C. Englert, R. S. Gupta, and M. Spannowsky, Probing Electroweak Precision Physics via boosted Higgs-strahlung at the LHC, Phys. Rev. D 98, 095012 (2018).
  71. S. Banerjee, R. S. Gupta, J. Y. Reiness, and M. Spannowsky, Resolving the tensor structure of the Higgs coupling to Z-bosons via Higgs-strahlung, Phys. Rev. D 100, 115004 (2019).
  72. G. Arcadi, A. Djouadi, and M. Raidal, Dark Matter through the Higgs portal, Phys. Rep. 842, 1 (2020).

  73. We envision dealing with Breit-Wigner resonances in two ways. For higher-order squared amplitudes, one can normalize by a lower order (or even the Born-level) amplitude—i.e., to produce a fully differential k factor. Another possibility is to approximate the amplitude itself, rather than its square, and decompose it into subamplitudes where the resonant propagators can be factored out. The need to address this case arises when considering four-lepton production via an off-shell diboson pair, and its detailed investigation is the subject of future work.

  74. We determined this evaluation time using openloops [75] to generate the exact gg→ZZ squared amplitudes for the training and prediction sets. The CPU core used for this timing is the same one used for all timings in this paper; see the main text for the exact specification.

  75. F. Buccioni, J.-N. Lang, J. M. Lindert, P. Maierhöfer, S. Pozzorini, H. Zhang, and M. F. Zoller, OpenLoops 2, Eur. Phys. J. C 79, 866 (2019).
  76. M. Abadi et al., tensorflow: Large-scale machine learning on heterogeneous systems (2015), software available from tensorflow.org.
  77. M. N. Wright and A. Ziegler, ranger: A fast implementation of random forests for high dimensional data in C++ and R, J. Stat. Softw. 77, 1 (2017).
  78. T. K. Ho, in Proceedings of the Third International Conference on Document Analysis and Recognition (Volume 1)—Volume 1, ICDAR ’95 (IEEE Computer Society, Washington, DC, USA, 1995), p. 278.
  79. L. Breiman, Random forests, Mach. Learn. 45, 5 (2001).
  80. L. Breiman, Arcing the edge, Technical Report 486, Statistics Department, University of California at Berkeley, 1997.
  81. J. H. Friedman, Greedy function approximation: A gradient boosting machine, Ann. Stat. 29, 1189 (2001).
  82. T. Chen and C. Guestrin, XGBoost, in Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining—KDD ’Z16 (2016), 10.1145/2939672.2939785.

  83. xgboost rose to prominence by winning the Kaggle Higgs Challenge. Since then, it has been consistently on the top of the ladder in a large number of the Kaggle competitions [84, 85], outperforming other ML architectures, like DNNs or SVMs.

  84. https://www.kaggle.com/sudalairajkumar/winning-solutions-of-kaggle-competitions.
  85. https://www.kaggle.com/rajiao/winning-solutions-of-kaggle-competitions.
  86. J. M. Campbell and R. K. Ellis, MCFM for the Tevatron and the LHC, Nucl. Phys. B, Proc. Suppl. 205–206, 10 (2010).
  87. M. Matsumoto and T. Nishimura, Mersenne twister: A 623-dimensionally equidistributed uniform pseudo-random number generator, ACM Trans. Model. Comput. Simul. 8, 3 (1998).

  88. We also tested madloops [2] but found it to be a factor of a few slower than openloops [75].

  89. E. Bothmann et al., Event generation with Sherpa 2.2, SciPost Phys. 7, 034 (2019).
  90. W. Kilian, S. Brass, T. Ohl, J. Reuter, V. Rothe, P. Stienemeier, and M. Utsch, New Developments in WHIZARD Version 2.6, in Proceedings of the International Workshop on Future Linear Collider (LCWS2017) Strasbourg, France, 2017 (2018), arXiv:1801.08034.

Outline

Information

Phys. Rev. D 107, L071901– Published 11 April, 2023
  • Received 16 April 2020
  • Revised 22 September 2020
  • Accepted 16 March 2023
Crossmark - Check for updates button
Creative Commons Logo

Published by the American Physical Society under the terms of the Creative Commons Attribution 4.0 International license. Further distribution of this work must maintain attribution to the author(s) and the published article’s title, journal citation, and DOI. Funded by SCOAP3.

Published by the American Physical Society

Sign In to Your Journals Account

Filter

Filter

Article Lookup

Enter a citation