- 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
  • Open Access

Mixture density networks for neutrino reconstruction at hadron colliders

Seungjin Yang

Jason S. H. Lee*

Junghwan Goh

  • *Contact author: jason.lee@yonsei.ac.kr
PDF

Phys. Rev. D 114, 076019 – Published 13 October, 2026

DOI: https://doi.org/10.1103/lz58-x4s8

Abstract

Neutrino momentum reconstruction at hadron colliders is intrinsically ambiguous because the longitudinal momentum is not directly observed. We study this problem in semileptonic tt¯ events using mixture of neutrino solutions with transformer event representation (monster), a mixture density network that predicts a multivariate normal mixture for the conditional distribution of neutrino momentum from reconstructed event objects. The model uses a transformer-based event encoder and yields a sampling-free point estimate of the neutrino momentum from the closed-form density. On the public benchmark introduced with ν-flows, monster reduces the 68th percentile of the three-momentum residual by 5% and the 95th percentile by 6% relative to the empirical-mode ν-flows baseline, with per-variable root-mean-square errors smaller by 2 to 9% at comparable bias, while running 2 to 8.5 times faster at inference, depending on the device and batch size. These results indicate that mixture density networks are a competitive alternative to normalizing flows for neutrino reconstruction.

View figure in article

Physics Subject Headings (PhySH)

Article Text

References (50)

  1. L. R. Evans and P. Bryant, LHC machine, J. Instrum. 3, S08001 (2008).
  2. M. Leigh, J. A. Raine, K. Zoch, and T. Golling, ν-flows: Conditional neutrino regression, SciPost Phys. 14, 159 (2023).
  3. G. Papamakarios, T. Pavlakou, and I. Murray, Masked autoregressive flow for density estimation, arXiv:1705.07057.
  4. D. J. Rezende and S. Mohamed, Variational Inference with Normalizing Flows, in Proceedings of the 32nd International Conference on Machine Learning, Proceedings of Machine Learning Research Vol. 37 (PMLR, Lille, France, 2015), pp. 1530–1538.
  5. J. A. Raine, M. Leigh, K. Zoch, and T. Golling, Fast and improved neutrino reconstruction in multineutrino final states with conditional normalizing flows, Phys. Rev. D 109, 012005 (2024).
  6. C. M. Bishop, Mixture density networks, Technical Report, NCRG/94/004, Aston University, 1994.
  7. G. Aad et al. (ATLAS Collaboration), The ATLAS experiment at the CERN Large Hadron Collider, J. Instrum. 3, S08003 (2008).
  8. E. E. Khoda (ATLAS Collaboration), ATLAS pixel cluster splitting using mixture density networks, Proc. Sci. LHCP2019 (2019) 009.
  9. ATLAS, Performance of ATLAS pixel detector and track reconstruction at the start of run 3 in LHC collisions at s=900  GeV, Technical Report, ATL-PHYS-PUB-2022-033, CERN, Geneva, 2022.
  10. ATLAS, Performance of mass-decorrelated jet substructure observables for hadronic two-body decay tagging in ATLAS, Technical Report ATL-PHYS-PUB-2018-014, CERN, Geneva, 2018.
  11. G. Graziani, L. Anderlini, S. Mariani, E. Franzoso, L. L. Pappalardo, and P. di Nezza, A neural-network-defined Gaussian mixture model for particle identification applied to the LHCb fixed-target programme, J. Instrum. 17, P02018 (2022).
  12. G. Aad et al. (ATLAS Collaboration), Simultaneous energy and mass calibration of large-radius jets with the ATLAS detector using a deep neural network, Mach. Learn. Sci. Tech. 5, 035051 (2024).
  13. A. R. Das et al., On-chip probabilistic inference for charged-particle tracking at the sensor edge, arXiv:2602.15946.
  14. K. Zoch, J. A. Raine, L. Ehrke, D. Sengupta, M. Leigh, and T. Golling, Semileptonic ttbar neutrino regression dataset, (2022).
  15. R. Xiong, Y. Yang, D. He, K. Zheng, S. Zheng, C. Xing, H. Zhang, Y. Lan, L. Wang, and T. Liu, On Layer Normalization in the Transformer Architecture, in Proceedings of the 37th International Conference on Machine Learning, ICML 2020, 13–18 July 2020, Virtual Event, Proceedings of Machine Learning Research Vol. 119 (PMLR, 2020), pp. 10524–10533, http://proceedings.mlr.press/v119/xiong20b.html.
  16. D. Hendrycks and K. Gimpel, Gaussian Error Linear Units (GELUs), arXiv:1606.08415.
  17. N. Srivastava, G. E. Hinton, A. Krizhevsky, I. Sutskever, and R. Salakhutdinov, Dropout: A simple way to prevent neural networks from overfitting, J. Mach. Learn. Res. 15, 1929 (2014).
  18. A. Dosovitskiy, L. Beyer, A. Kolesnikov, D. Weissenborn, X. Zhai, T. Unterthiner, M. Dehghani, M. Minderer, G. Heigold, S. Gelly, J. Uszkoreit, and N. Houlsby, An image is worth 16×16 words: Transformers for image recognition at scale, arXiv:2010.11929.
  19. C. Dugas, Y. Bengio, F. Bélisle, C. Nadeau, and R. Garcia, Incorporating second-order functional knowledge for better option pricing, in Advances in Neural Information Processing Systems, edited by T. Leen, T. Dietterich, and V. Tresp (MIT Press, Cambridge, MA, 2000), Vol. 13.
  20. D. C. Liu and J. Nocedal, On the limited memory BFGS method for large scale optimization, Math. Program. 45, 503 (1989).
  21. P. Wolfe, Convergence conditions for ascent methods, SIAM Rev. 11, 226 (1969).
  22. P. Wolfe, Convergence conditions for ascent methods. II: Some corrections, SIAM Rev. 13, 185 (1971).
  23. J. Nocedal and S. J. Wright, Numerical Optimization, 2nd ed., Springer Series in Operations Research and Financial Engineering (Springer, New York, 2006).
  24. I. Loshchilov and F. Hutter, Decoupled weight decay regularization, arXiv:1711.05101.
  25. P. Goyal, P. Dollár, R. Girshick, P. Noordhuis, L. Wesolowski, A. Kyrola, A. Tulloch, Y. Jia, and K. He, Accurate, large minibatch SGD: Training ImageNet in 1 hour, arXiv:1706.02677.
  26. I. Loshchilov and F. Hutter, SGDR: Stochastic gradient descent with warm restarts, arXiv:1608.03983.
  27. 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.
  28. P. Artoisenet, R. Frederix, O. Mattelaer, and R. Rietkerk, Automatic spin-entangled decays of heavy resonances in Monte Carlo simulations, J. High Energy Phys. 03 (2013) 015.
  29. T. Sjöstrand, S. Mrenna, and P. Skands, A brief introduction to PYTHIA 8.1, Comput. Phys. Commun. 178, 852 (2008).
  30. J. de Favereau, C. Delaere, P. Demin, A. Giammanco, V. Lemaître, A. Mertens, and M. Selvaggi (DELPHES 3 Collaboration), DELPHES 3, A modular framework for fast simulation of a generic collider experiment, J. High Energy Phys. 02 (2014) 057.
  31. M. Cacciari, G. P. Salam, and G. Soyez, The anti-kt jet clustering algorithm, J. High Energy Phys. 04 (2008) 063.
  32. L. Ardizzone, C. Lüth, J. Kruse, C. Rother, and U. Köthe, Guided image generation with conditional invertible neural networks, arXiv:1907.02392.
  33. M. Zaheer, S. Kottur, S. Ravanbakhsh, B. Póczos, R. Salakhutdinov, and A. Smola, Deep sets, arXiv:1703.06114.
  34. C. Durkan, A. Bekasov, I. Murray, and G. Papamakarios, Neural spline flows, in Advances in Neural Information Processing Systems (Curran Associates, Inc., Red Hook, NY, 2019), Vol. 32.
  35. D. P. Kingma and P. Dhariwal, Glow: Generative flow with invertible 1×1 convolutions, in Advances in Neural Information Processing Systems (Curran Associates, Inc., Red Hook, NY, 2018), Vol. 31.
  36. B. Xu, N. Wang, T. Chen, and M. Li, Empirical evaluation of rectified activations in convolutional network, arXiv:1505.00853.
  37. J. L. Ba, J. R. Kiros, and G. E. Hinton, Layer normalization arXiv:1607.06450.
  38. K. He, X. Zhang, S. Ren, and J. Sun, Deep residual learning for image recognition, 10.1109/CVPR.2016.90 (2015).
  39. K. He, X. Zhang, S. Ren, and J. Sun, Identity mappings in deep residual networks, arXiv:1603.05027.
  40. M. Leigh, neutrino_flows, https://github.com/mattcleigh/neutrino_flows (2022) (accessed 2026-03-30).
  41. G. Aad et al. (ATLAS Collaboration), Determination of the top-quark pole mass using tt¯+1-jet events collected with the ATLAS experiment in 7 TeV pp collisions, J. High Energy Phys. 10 (2015) 121.
  42. A. M. Sirunyan et al. (CMS Collaboration), Measurement of the top quark mass with lepton+jets final states using pp collisions at s=13  TeV, Eur. Phys. J. C 78, 891 (2018); 82, 323(E) (2022).
  43. J. Kvita, Study of methods of resolved top quark reconstruction in semileptonic tt¯ decay, Nucl. Instrum. Methods Phys. Res., Sect. A 900, 84 (2018); 1040, 167172(E) (2022).
  44. M. Aaboud et al. (ATLAS Collaboration), Measurement of the top quark mass in the tt¯→lepton+jets channel from s=8  TeV ATLAS data and combination with previous results, Eur. Phys. J. C 79, 290 (2019).
  45. G. Aad et al. (ATLAS Collaboration), Measurement of the top-quark mass in tt¯+1-jet events collected with the ATLAS detector in pp collisions at s=8  TeV, J. High Energy Phys. 11 (2019) 150.
  46. G. Aad et al. (ATLAS Collaboration), Measurements of top-quark pair differential and double-differential cross sections in the ℓ+jets channel with pp collisions at s=13  TeV using the ATLAS detector, Eur. Phys. J. C 79, 1028 (2019); 80, 1092(E) (2020).
  47. M. Grossi, J. Novak, B. Kersevan, and D. Rebuzzi, Comparing traditional and deep-learning techniques of kinematic reconstruction for polarization discrimination in vector boson scattering, Eur. Phys. J. C 80, 1144 (2020).
  48. A. Tumasyan et al. (CMS Collaboration), Measurement of differential tt¯ production cross sections in the full kinematic range using lepton+jets events from proton–proton collisions at s=13  TeV, Phys. Rev. D 104, 092013 (2021).
  49. G. Aad et al. (ATLAS Collaboration), Evidence for the charge asymmetry in pp→tt¯ production at s=13  TeV with the ATLAS detector, J. High Energy Phys. 08 (2023) 077.
  50. S. Yang, MoNSTER, GitHub (2026), https://github.com/slowmoyang/monster.

Outline

Information

Phys. Rev. D 114, 076019– Published 13 October, 2026
  • Received 22 May 2026
  • Accepted 18 September 2026
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