1932

Abstract

Machine learning has played an important role in the analysis of high-energy physics data for decades. The emergence of deep learning in 2012 allowed for machine learning tools which could adeptly handle higher-dimensional and more complex problems than previously feasible. This review is aimed at the reader who is familiar with high-energy physics but not machine learning. The connections between machine learning and high-energy physics data analysis are explored, followed by an introduction to the core concepts of neural networks, examples of the key results demonstrating the power of deep learning for analysis of LHC data, and discussion of future prospects and concerns.

Loading

Article metrics loading...

/content/journals/10.1146/annurev-nucl-101917-021019
2018-10-19
2025-02-11
Loading full text...

Full text loading...

/deliver/fulltext/nucl/68/1/annurev-nucl-101917-021019.html?itemId=/content/journals/10.1146/annurev-nucl-101917-021019&mimeType=html&fmt=ahah

Literature Cited

  1. 1.  Hocker A et al. Proc. Sci. ACAT:040 2007.
    [Google Scholar]
  2. 2.  Krizhevsky A, Sutskever I, Hinton GEIn Proceedings of the 25th International Conference on Neural Information Processing Systems (NIPS 12)1097 New York: Curran 2012.
    [Google Scholar]
  3. 3.  Russakovsky O et al. Int. J. Comput. Vis. 115:211 2015.
    [Google Scholar]
  4. 4.  LeCun Y, Bengio Y, Hinton G Nature 521:436 2015.
    [Google Scholar]
  5. 5.  Schmidhuber J Neural Netw. 61:85 2015.
    [Google Scholar]
  6. 6.  Sjöstrand T, Mrenna S, Skands PZ J. High Energy Phys. 0605:026 2006.
    [Google Scholar]
  7. 7.  Bahr M et al. Eur. Phys. J. C 58:639 2008.
    [Google Scholar]
  8. 8.  Alwall J et al. J. High Energy Phys. 07:079 2014.
    [Google Scholar]
  9. 9.  Gleisberg T et al. J. High Energy Phys. 02:007 2009.
    [Google Scholar]
  10. 10.  Agostinelli S et al. Nucl. Instrum. Methods A 506:250 2003.
    [Google Scholar]
  11. 11. Keras Document. Team. Keras Software package https://keras.io/ 2018.
    [Google Scholar]
  12. 12. LISA Lab. Theano Software package http://deeplearning.net/software/theano/ 2018.
    [Google Scholar]
  13. 13. TensorFlow Dev. Team. TensorFlow Software package https://www.tensorflow.org/ 2018.
    [Google Scholar]
  14. 14.  Hornik K, Stinchcombe M, White H Neural Netw. 2:359 1989.
    [Google Scholar]
  15. 15.  Hochreiter S Int. J. Uncertain. Fuzziness Knowl.-Based Syst. 6:107–116 1998.
    [Google Scholar]
  16. 16.  Bengio Y, Simard P, Frasconi P Neural Netw. 5:157 1994.
    [Google Scholar]
  17. 17.  Hinton G et al. arXiv:1207.0580 [cs.NE] 2012.
  18. 18.  Hinton GE, Osindero S, Teh YW Neural Comput. 18:1527 2006.
    [Google Scholar]
  19. 19.  Bengio Y, Lamblin P, Popovici D, Larochelle H Proceedings of the 19th Annual Conference on Advances in Neural Information Processing Systems (NIPS 06) B Schölkopf, J Platt, T Hoffman153 Cambridge, MA: MIT Press 2007.
    [Google Scholar]
  20. 20.  Goodfellow I, Bengio Y, Courville A Deep Learning Cambridge, MA: MIT Press http://www.deeplearningbook.org 2016.
    [Google Scholar]
  21. 21.  Lecun Y, Bottou L, Bengio Y, Haffner P Proc. IEEE 86:2278 1998.
    [Google Scholar]
  22. 22.  Ioffe S, Szegedy C Proceedings of the 32nd International Conference on Machine Learning448 proceedings.mlr.press: PMLR 2015.
    [Google Scholar]
  23. 23.  He K, Zhang X, Ren S, Sun J Proceedings of the IEEE Conference on Computer Vision and Pattern Recognitionp 1026 Piscataway, NJ: IEEE 2016.
    [Google Scholar]
  24. 24.  Thulasiraman K, Swamy MNS Graphs: Theory and Algorithms New York: Wiley 1992.
    [Google Scholar]
  25. 25.  Hochreiter S, Schmidhuber J Neural Comput. 9:1735 1997.
    [Google Scholar]
  26. 26.  Cho K et al. arXiv:1406.1078 [cs] 2014.
  27. 27.  Aad G et al. J. Instrum. 9:P09009 2014.
    [Google Scholar]
  28. 28.  Peterson C Nucl. Instrum. Methods A 279:537 1989.
    [Google Scholar]
  29. 29.  Abramowicz H, Caldwell A, Sinkus R Nucl. Instrum. Methods A 365:508 1995.
    [Google Scholar]
  30. 30. CMS Collab. Eur. Phys. J. C 74:3076 2014.
    [Google Scholar]
  31. 31.  Abazov VM et al. Phys. Rev. D 71:072004 2005.Erratum. Phys. Rev. D 77:039901 2008.
    [Google Scholar]
  32. 32. DELPHI Collab. Phys. Lett. B 295:383 1992.
    [Google Scholar]
  33. 33.  Kohne JK et al. Nucl. Instrum. Methods A 389:128 1997.
    [Google Scholar]
  34. 34. D0 Collab. Phys. Lett. B 517:282 2001.
    [Google Scholar]
  35. 35. CDF Collab. Phys. Rev. Lett. 104:141801 2010.
    [Google Scholar]
  36. 36.  Baldi P, Sadowski P, Whiteson D Nat. Commun. 5:4308 2014.
    [Google Scholar]
  37. 37.  Santos R et al. J. Instrum. 12:P04014 2017.
    [Google Scholar]
  38. 38.  Louppe G, Cranmer K, Pavez J J. Open Source Softw. 1:11 2016.
    [Google Scholar]
  39. 39.  Baldi P et al. Eur. Phys. J. C 76:235 2016.
    [Google Scholar]
  40. 40.  Behnke T, Charlton DG Phys. Scr. 52:133 1995.
    [Google Scholar]
  41. 41.  Peterson C, Rögnvaldsson T, Lönnblad L Comput. Phys. Commun. 81:185 1994.
    [Google Scholar]
  42. 42. CMS Collab. Tech. rep. CMS-PAS-BTV-15-001, CERN, Geneva 2016.
  43. 43. ATLAS Collab. J. Instrum. 11:P04008 2016.
    [Google Scholar]
  44. 44. ATLAS Collab. Tech. rep. ATL-PHYS-PUB-2017-013, CERN, Geneva 2017.
  45. 45.  Larkoski AJ, Moult I, Nachman B arXiv:1709.04464 [hep-ph] 2017.
  46. 46.  de Oliveira L et al. J. High Energy Phys. 07:069 2016.
    [Google Scholar]
  47. 47.  Cogan J, Kagan M, Strauss E, Schwarztman A J. High Energy Phys. 02:118 2015.
    [Google Scholar]
  48. 48.  Almeida LG et al. J. High Energy Phys. 07:086 2015.
    [Google Scholar]
  49. 49.  Baldi P et al. Phys. Rev. D 93:094034 2016.
    [Google Scholar]
  50. 50.  Barnard J, Dawe EN, Dolan MJ, Rajcic N Phys. Rev. D 95:014018 2017.
    [Google Scholar]
  51. 51.  Komiske PT, Metodiev EM, Schwartz MD J. High Energy Phys. 01:110 2017.
    [Google Scholar]
  52. 52.  Acciarri R et al. J. Instrum. 12:P03011 2017.
    [Google Scholar]
  53. 53.  Aurisano A et al. J. Instrum. 11:P09001 2016.
    [Google Scholar]
  54. 54.  Guest D et al. Phys. Rev. D 94:112002 2016.
    [Google Scholar]
  55. 55.  CMS Collab J. Instrum. 12:P10003 2017.
    [Google Scholar]
  56. 56. ATLAS Collab. Tech. rep. ATL-PHYS-PUB-2017-003, CERN, Geneva 2017.
  57. 57. CMS Collab. Tech. rep. CMS-DP-2017-012, CERN, Geneva 2017.
  58. 58. CMS Collab. Tech. rep. CMS-DP-2017-005, CERN, Geneva 2017.
  59. 59. CMS Collab. Tech. rep. CMS-DP-2017-049, CERN, Geneva 2017.
  60. 60.  Pearkes J, Fedorko W, Lister A, Gay C arXiv:1704.02124 [hep-ex] 2017.
  61. 61.  Egan S et al. arXiv:1711.09059 [hep-ex] 2017.
  62. 62. CMS Collab. Presented at the Workshop on Deep Learning for Physical Sciences of the 31st Annual Conference on Advances in Neural Information Processing Systems (NIPS 17), Long Beach, CA 2017.
  63. 63.  Louppe G, Cho K, Becot C, Cranmer K arXiv:1702.00748 [hep-ph] 2017.
  64. 64.  Cheng T arXiv:1711.02633 [hep-ph] 2017.
  65. 65.  Henrion I et al. Presented at the Workshop on Deep Learning for Physical Sciences of the 31st Annual Conference on Advances in Neural Information Processing Systems (NIPS 17), Long Beach, CA 2017.
  66. 66.  Bronstein MM et al. arXiv:1611.08097 [cs.CV] 2016.
  67. 67. ATLAS Collab. Eur. Phys. J. C 77:673 2017.
    [Google Scholar]
  68. 68.  Salzburger A Tech. rep. ATL-SOFT-PROC-2015-056.7, CERN, Geneva 2015.
  69. 69.  Stahl M J. Phys. Conf. Ser. 898:042042 2017.
    [Google Scholar]
  70. 70.  Machado Miguens J Tech. rep. ATL-DAQ-PROC-2016-025, CERN, Geneva 2016.
  71. 71.  Adam W, Fruhwirth R, Strandlie A, Todorov T J. Phys. G 31:9 2005.
    [Google Scholar]
  72. 72. ATLAS Collab. Tech. rep. ATLAS-CONF-2012-047, CERN, Geneva 2012.
  73. 73.  Salzburger A J. Phys. Conf. Ser. 664:072042 2015.
    [Google Scholar]
  74. 74.  Krishnan RG, Shalit U, Sontag D. arXiv:1511.05121 [stat.ML] 2015.
  75. 75.  Krishnan RG, Shalit U, Sontag D Proceedings of the 31st AAAI Conference on Artificial Intelligencep 2101 Menlo Park, CA: AAAI 2017.
    [Google Scholar]
  76. 76.  Farrell S et al. Eur. Phys. J. Web Conf. 150:00003 2017.
    [Google Scholar]
  77. 77.  Farrell S et al. Presented at the Workshop on Deep Learning for Physical Sciences of the 31st Annual Conference on Advances in Neural Information Processing Systems (NIPS 17), Long Beach, CA. https://dl4physicalsciences.github.io/files/nips_dlps_2017_28.pdf 2017.
  78. 78.  Amrouche S et al. Eur. Phys. J. Web Conf. 150:00015 2017.
    [Google Scholar]
  79. 79. ATLAS Collab. Tech. rep. ATL-PHYS-PUB-2010-013, ATL-COM-PHYS-2010-838, CERN, Geneva 2010.
  80. 80.  Larochelle H, Murray I Proceedings of the 14th International Conference on Artificial Intelligence and Statisticsp 29 Cambridge, MA: MIT Press 2011.
    [Google Scholar]
  81. 81.  Jimenez Rezende D, Mohamed S. arXiv:1505.05770 [stat.ML] 2015.
  82. 82.  Dinh L, Krueger D, Bengio Y arXiv:1410.8516 [cs.LG] 2014.
  83. 83.  Cranmer K, Pavez J, Louppe G arXiv:1506.02169 [stat.AP] 2015.
  84. 84.  Kingma DP, Salimans T, Welling M arXiv:1606.04934 2016.
  85. 85.  Papamakarios G, Murray I, Pavlakou T Proceedings of the 31st Annual Conference on Advances in Neural Information Processing Systems (NIPS 17)2335 New York: Curran
    [Google Scholar]
  86. 86.  Paganini M, de Oliveira L, Nachman B arXiv:1705.02355 [hep-ex] 2017.
  87. 87.  de Oliveira L, Paganini M, Nachman B arXiv:1701.05927 [stat.ML] 2017.
  88. 88.  Elmer P, Neubauer M, Sokoloff MD arXiv:1712.06592 [physics.comp-ph] 2017.
  89. 89.  Alves AA Jr. et al. arXiv:1712.06982 [physics.comp-ph] 2017.
  90. 90.  Aaltonen T et al. Phys. Rev. Lett. 102:152001 2009.
    [Google Scholar]
  91. 91.  Stevens J, Williams M J. Instrum. 8:P12013 2013.
    [Google Scholar]
  92. 92.  Ajakan H et al. arXiv:1412.4446 [stat.ML] 2014.
  93. 93.  Louppe G, Kagan M, Cranmer K arXiv:1611.01046 [stat.ME] 2016.
  94. 94.  Shimmin C et al. arXiv:1703.03507 [hep-ex] 2017.
  95. 95.  Snoek J, Larochelle H, Adams RP. arXiv:1206.2944 [stat.ML] 2012.
  96. 96.  Head T et al. Scikit-Optimize. Software package https://scikit-optimize.github.io/ 2017.
    [Google Scholar]
  97. 97.  Staines J, Barber D. arXiv:1212.4507 [stat] 2012.
  98. 98.  Louppe G, Cranmer K. arXiv:1707.07113 [stat] 2017.
  99. 99.  Chang S, Cohen T, Ostdiek B arXiv:1709.10106 [hep-ph] 2017.
  100. 100.  Cranmer K, Yavin I J. High Energy Phys. 04:038 2011.
    [Google Scholar]
  101. 101.  Cranmer K, Heinrich L J. Phys. Conf. Ser. 898:102019 2017.
    [Google Scholar]
  102. 102.  Dery LM, Nachman B, Rubbo F, Schwartzman A J. High Energy Phys. 05:145 2017.
    [Google Scholar]
  103. 103.  Metodiev EM, Nachman B, Thaler J J. High Energy Phys. 10:174 2017.
    [Google Scholar]
  104. 104. ONNX Dev. Team. Open Neural Network Exchange (ONNX). Software package https://github.com/onnx/onnx 2017.
    [Google Scholar]
  105. 105.  Guest DH et al. lwtnn (version 2.6) Software package https://github.com/lwtnn/lwtnn/releases 2017.
    [Google Scholar]
  106. 106.  Tran D et al. Edward. Software package http://edwardlib.org 2017.
    [Google Scholar]
  107. 107.  Tran D et al. arXiv:1610.09787 [stat] 2016.
  108. 108.  Pyro Dev. Team. Pyro Software package http://pyro.ai 2017.
    [Google Scholar]
  109. 109.  Frate M et al. arXiv:1709.05681 [physics.data-an] 2017.
  110. 110.  Luo C, Shrivastava A arXiv:1706.06664 [cs.DB] 2017.
  111. 111.  Ilten P, Williams M, Yang Y J. Instrum. 12:P04028 2017.
    [Google Scholar]
  112. 112.  Bertone G et al. arXiv:1611.02704 [hep-ph] 2016.
  113. 113.  Caron S et al. Eur. Phys. J. C 77:257 2017.
    [Google Scholar]
  114. 114.  Cranmer K, Louppe G J. Brief Ideas http://doi.org10.5281/zenodo.198541 2016.
    [Google Scholar]
  115. 115.  Le TA, Baydin AG, Wood F arXiv:1610.09900 [cs.AI] 2016.
  116. 116.  Casado ML et al. Presented at the Workshop on Deep Learning for Physical Sciences of the 31st Annual Conference on Advances in Neural Information Processing Systems (NIPS 17), Long Beach, CA 2017.
/content/journals/10.1146/annurev-nucl-101917-021019
Loading
/content/journals/10.1146/annurev-nucl-101917-021019
Loading

Data & Media loading...

  • Article Type: Review Article
This is a required field
Please enter a valid email address
Approval was a Success
Invalid data
An Error Occurred
Approval was partially successful, following selected items could not be processed due to error