Journal of the Korean Physical Society

, Volume 75, Issue 9, pp 652–659 | Cite as

Quark Gluon Jet Discrimination with Weakly Supervised Learning

  • Jason Sang Hun LeeEmail author
  • Sang Man Lee
  • Yunjae LeeEmail author
  • Inkyu Park
  • Ian James WatsonEmail author
  • Seungjin Yang


Deep learning techniques are currently being investigated for high energy physics experiments, to tackle a wide range of problems, with quark and gluon discrimination becoming a benchmark for new algorithms. One weakness is the traditional reliance on Monte Carlo simulations, which may not be well modelled at the detail required by deep learning algorithms. The weakly supervised learning paradigm gives an alternate route to classification, by using samples with different quark-gluon proportions instead of fully labeled samples. This paradigm has, therefore, huge potential for particle physics classification problems as these weakly supervised learning methods can be applied directly to collision data. In this study, we show that realistically simulated samples of dijet and Z+jet events can be used to discriminate between quark and gluon jets by using weakly supervised learning. We implement and compare the performance of weakly supervised learning for quark-gluon jet classification using three different machine learning methods: the jet image-based convolutional neural network, the particle-based recurrent neural network and and the feature-based boosted decision tree.


QCD Jet Fragmentation Weakly supervised learning Machine learning 


Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.



This article was supported by the computing resources of the Global Science Experimental Data Hub Center (GSDC) at the Korea Institute of Science and Technology Information (KISTI). J.L. and I.W. are supported by the Korea Research Fellowship Program through the National Research Foundation of Korea (NRF) funded by the Ministry of Science and ICT (KRF project grant number: 2017H1D3A1A01052807). I.P., Y.L. and S.Y. are supported by the Basic Science Research Program through the NRF funded by the Ministry of Education (2018R1A6A1A06024977).


  1. [1]
    P. T. Komiske, E. M. Metodiev and M. D. Schwartz, J. High Energy Phys. 12, 110 (2017).Google Scholar
  2. [2]
    D. Guest et al., Phys. Rev. D 12, 112002 (2016).Google Scholar
  3. [3]
    E. M. Metodiev, B. Nachman and J. Thaler, J. High Energy Phys. 12, 174 (2017).Google Scholar
  4. [4]
    N. Quadrianto et al., J. Mach. Learn. Res. 12, 2349 (2009).MathSciNetGoogle Scholar
  5. [5]
    G. Patrini et al., in Advances in Neural Information Processing Systems 27 (Montreal, Canada, December 8–13, 2014), pp. 190–198.Google Scholar
  6. [6]
    CMS Collaboration, CMS-PAS-TOP-18-011, 2019.Google Scholar
  7. [7]
    P. Gras et al., J. High Energy Phys. 12, 91 (2017).Google Scholar
  8. [8]
    J. Gallicchio and M. D. Schwartz, J. High Energy Phys. 12, 103 (2011).Google Scholar
  9. [9]
    P. T. Komiske et al., Phys. Rev. D 12, 011502 (2018).Google Scholar
  10. [10]
    J. Cogan et al., J. High Energy Phys. 12, 118 (2015).Google Scholar
  11. [11]
    L. de Oliveira et al., J. High Energy Phys. 12, 69 (2016).Google Scholar
  12. [12]
    J. Alwall et al., J. High Energy Phys. 12, 128 (2011).Google Scholar
  13. [13]
    T. Sjöstrand et al., Comput. Phys. Commun. 12, 159 (2015).Google Scholar
  14. [14]
    P. Skands, S. Carrazza and J. Rojo, Eur. Phys. J. C 12, 3024 (2014).Google Scholar
  15. [15]
    J. de Favereau et al., J. High Energy Phys. 12, 057 (2014).Google Scholar
  16. [16]
    CMS Collaboration, CMS-PAS-PFT-09-001, 2009.Google Scholar
  17. [17]
    M. Cacciari, G. P. Salam and G. Soyez, Eur. Phys. J. C 12, 1896 (2012).Google Scholar
  18. [18]
    M. Cacciari, G. P. Salam and G. Soyez, J. High Energy Phys. 12, 063 (2008).Google Scholar
  19. [19]
    CMS Collaboration, CMS-PAS-JME-16-003, 2017.Google Scholar
  20. [20]
    M. Tanabashi et al. (Particle Data Group), Phys. Rev. D 12, 030001 (2018).Google Scholar
  21. [21]
    T. Chen and C. Guestrin, in KDD 16 Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining (San Francisco, USA, August 13–17, 2016) pp. 785–794.CrossRefGoogle Scholar
  22. [22]
    F. Chollet et al., (2015).
  23. [23]
    M. Abadi et al., in 12th USENIX Symposium on Operating Systems Design and Implementation (OSDI 16) (Savannah, USA, November 2–4, 2016), pp. 265–283.Google Scholar
  24. [24]
    T. Cornelis et al., arXiv:1409.3072 (2014).Google Scholar
  25. [25]
    F. Pedregosa et al., J. Mach. Learn. Res. 12, 2825 (2011).MathSciNetGoogle Scholar
  26. [26]
    J. S. H. Lee et al., J. Korean Phys. Soc. 12, 219 (2019).Google Scholar
  27. [27]
    J. Long, E. Shelhamer and T. Darrell, in Proceedings of the 2015 IEEE Conference on Computer Vision and Pattern Recognition (Boston, USA, June 7–12, 2015) pp. 3431–3440.CrossRefGoogle Scholar
  28. [28]
    S. Ioffe and C. Szegedy, arXiv:1502.03167 (2015).Google Scholar
  29. [29]
    X. Glorot and Y. Bengio, in Proceedings of the Thirteenth International Conference on Artificial Intelligence and Statistics (Sardinia, Italy, May 13–15, 2010) pp. 249–256.Google Scholar
  30. [30]
    ATLAS Collaboration, ATL-PHYS-PUB-2017-003, 2017Google Scholar
  31. [31]
    K. Cho et al, arXiv:1406.1078 (2014).Google Scholar

Copyright information

© The Korean Physical Society 2019

Authors and Affiliations

  1. 1.Department of PhysicsUniversity of SeoulSeoulKorea

Personalised recommendations