Papers
Topics
Authors
Recent
Gemini 2.5 Flash
Gemini 2.5 Flash
173 tokens/sec
GPT-4o
7 tokens/sec
Gemini 2.5 Pro Pro
46 tokens/sec
o3 Pro
4 tokens/sec
GPT-4.1 Pro
38 tokens/sec
DeepSeek R1 via Azure Pro
28 tokens/sec
2000 character limit reached

Normalizing Flows for High-Dimensional Detector Simulations (2312.09290v2)

Published 14 Dec 2023 in hep-ph

Abstract: Whenever invertible generative networks are needed for LHC physics, normalizing flows show excellent performance. In this work, we investigate their performance for fast calorimeter shower simulations with increasing phase space dimension. We use fast and expressive coupling spline transformations applied to the CaloChallenge datasets. In addition to the base flow architecture we also employ a VAE to compress the dimensionality and train a generative network in the latent space. We evaluate our networks on several metrics, including high-level features, classifiers, and generation timing. Our findings demonstrate that invertible neural networks have competitive performance when compared to autoregressive flows, while being substantially faster during generation.

Definition Search Book Streamline Icon: https://streamlinehq.com
References (114)
  1. J. M. Campbell et al., Event Generators for High-Energy Physics Experiments, In Snowmass 2021 (2022), arXiv:2203.11110.
  2. S. Badger et al., Machine learning and LHC event generation, SciPost Phys. 14(4), 079 (2023), 10.21468/SciPostPhys.14.4.079, arXiv:2203.07460.
  3. S. Agostinelli et al., GEANT4–a simulation toolkit, Nucl. Instrum. Meth. A 506, 250 (2003), 10.1016/S0168-9002(03)01368-8.
  4. Geant4 developments and applications, IEEE Transactions on Nuclear Science 53(1), 270 (2006), 10.1109/TNS.2006.869826.
  5. Recent developments in geant4, Nuclear Instruments and Methods in Physics Research Section A: Accelerators, Spectrometers, Detectors and Associated Equipment 835, 186 (2016), https://doi.org/10.1016/j.nima.2016.06.125, https://www.sciencedirect.com/science/article/pii/S0168900216306957.
  6. Modern Machine Learning for LHC Physicists (2022), arXiv:2211.01421.
  7. GANplifying event samples, SciPost Phys. 10(6), 139 (2021), 10.21468/SciPostPhys.10.6.139, arXiv:2008.06545.
  8. Calomplification — the power of generative calorimeter models, JINST 17(09), P09028 (2022), 10.1088/1748-0221/17/09/P09028, arXiv:2202.07352.
  9. J. Bendavid, Efficient Monte Carlo Integration Using Boosted Decision Trees and Generative Deep Neural Networks (2017), arXiv:1707.00028.
  10. M. D. Klimek and M. Perelstein, Neural Network-Based Approach to Phase Space Integration, SciPost Phys. 9, 053 (2020), 10.21468/SciPostPhys.9.4.053, arXiv:1810.11509.
  11. I.-K. Chen, M. D. Klimek and M. Perelstein, Improved Neural Network Monte Carlo Simulation, SciPost Phys. 10, 023 (2021), 10.21468/SciPostPhys.10.1.023, arXiv:2009.07819.
  12. C. Gao, J. Isaacson and C. Krause, i-flow: High-dimensional Integration and Sampling with Normalizing Flows, Mach. Learn. Sci. Tech. 1(4), 045023 (2020), 10.1088/2632-2153/abab62, arXiv:2001.05486.
  13. Exploring phase space with Neural Importance Sampling, SciPost Phys. 8(4), 069 (2020), 10.21468/SciPostPhys.8.4.069, arXiv:2001.05478.
  14. Event Generation with Normalizing Flows, Phys. Rev. D 101(7), 076002 (2020), 10.1103/PhysRevD.101.076002, arXiv:2001.10028.
  15. Accelerating Monte Carlo event generation – rejection sampling using neural network event-weight estimates, SciPost Phys. 12(5), 164 (2022), 10.21468/SciPostPhys.12.5.164, arXiv:2109.11964.
  16. MadNIS – Neural Multi-Channel Importance Sampling (2022), arXiv:2212.06172.
  17. Unweighting multijet event generation using factorisation-aware neural networks, SciPost Phys. 15, 107 (2023), 10.21468/SciPostPhys.15.3.107, arXiv:2301.13562.
  18. Efficient phase-space generation for hadron collider event simulation, SciPost Phys. 15, 169 (2023), 10.21468/SciPostPhys.15.4.169, arXiv:2302.10449.
  19. The MadNIS Reloaded (2023), arXiv:2311.01548.
  20. L. de Oliveira, M. Paganini and B. Nachman, Learning Particle Physics by Example: Location-Aware Generative Adversarial Networks for Physics Synthesis, Comput. Softw. Big Sci. 1(1), 4 (2017), 10.1007/s41781-017-0004-6, arXiv:1701.05927.
  21. JUNIPR: a Framework for Unsupervised Machine Learning in Particle Physics, Eur. Phys. J. C 79(2), 102 (2019), 10.1140/epjc/s10052-019-6607-9, arXiv:1804.09720.
  22. E. Bothmann and L. Debbio, Reweighting a parton shower using a neural network: the final-state case, JHEP 01, 033 (2019), 10.1007/JHEP01(2019)033, arXiv:1808.07802.
  23. K. Dohi, Variational Autoencoders for Jet Simulation (2020), arXiv:2009.04842.
  24. E. Buhmann, G. Kasieczka and J. Thaler, EPiC-GAN: Equivariant Point Cloud Generation for Particle Jets (2023), arXiv:2301.08128.
  25. PC-JeDi: Diffusion for Particle Cloud Generation in High Energy Physics (2023), arXiv:2303.05376.
  26. V. Mikuni, B. Nachman and M. Pettee, Fast Point Cloud Generation with Diffusion Models in High Energy Physics (2023), arXiv:2304.01266.
  27. EPiC-ly Fast Particle Cloud Generation with Flow-Matching and Diffusion (2023), arXiv:2310.00049.
  28. Modeling hadronization using machine learning, SciPost Phys. 14, 027 (2023), 10.21468/SciPostPhys.14.3.027, arXiv:2203.04983.
  29. Towards a deep learning model for hadronization, Phys. Rev. D 106(9), 096020 (2022), 10.1103/PhysRevD.106.096020, arXiv:2203.12660.
  30. Fitting a deep generative hadronization model, JHEP 09, 084 (2023), 10.1007/JHEP09(2023)084, arXiv:2305.17169.
  31. Towards a data-driven model of hadronization using normalizing flows (2023), arXiv:2311.09296.
  32. 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(4), 042003 (2018), 10.1103/PhysRevLett.120.042003, arXiv:1705.02355.
  33. L. de Oliveira, M. Paganini and B. Nachman, Controlling Physical Attributes in GAN-Accelerated Simulation of Electromagnetic Calorimeters, J. Phys. Conf. Ser. 1085(4), 042017 (2018), 10.1088/1742-6596/1085/4/042017, arXiv:1711.08813.
  34. 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(1), 014021 (2018), 10.1103/PhysRevD.97.014021, arXiv:1712.10321.
  35. Generating and refining particle detector simulations using the Wasserstein distance in adversarial networks, Comput. Softw. Big Sci. 2(1), 4 (2018), 10.1007/s41781-018-0008-x, arXiv:1802.03325.
  36. M. Erdmann, J. Glombitza and T. Quast, Precise simulation of electromagnetic calorimeter showers using a Wasserstein Generative Adversarial Network, Comput. Softw. Big Sci. 3(1), 4 (2019), 10.1007/s41781-018-0019-7, arXiv:1807.01954.
  37. D. Belayneh et al., Calorimetry with Deep Learning: Particle Simulation and Reconstruction for Collider Physics, Eur. Phys. J. C 80(7), 688 (2020), 10.1140/epjc/s10052-020-8251-9, arXiv:1912.06794.
  38. Getting High: High Fidelity Simulation of High Granularity Calorimeters with High Speed, Comput. Softw. Big Sci. 5(1), 13 (2021), 10.1007/s41781-021-00056-0, arXiv:2005.05334.
  39. Fast simulation of the ATLAS calorimeter system with Generative Adversarial Networks, Tech. Rep. ATL-SOFT-PUB-2020-006, CERN, Geneva, All figures including auxiliary figures are available at https://atlas.web.cern.ch/Atlas/GROUPS/PHYSICS/PUBNOTES/ATL-SOFT-PUB-2020-006 (2020).
  40. Decoding Photons: Physics in the Latent Space of a BIB-AE Generative Network (2021), arXiv:2102.12491.
  41. C. Krause and D. Shih, CaloFlow: Fast and Accurate Generation of Calorimeter Showers with Normalizing Flows, Phys. Rev. D 107(11), 113003 (2023), 10.1103/PhysRevD.107.113003, arXiv:2106.05285.
  42. G. Aad et al., AtlFast3: the next generation of fast simulation in ATLAS, Comput. Softw. Big Sci. 6, 7 (2022), 10.1007/s41781-021-00079-7, arXiv:2109.02551.
  43. C. Krause and D. Shih, CaloFlow II: Even Faster and Still Accurate Generation of Calorimeter Showers with Normalizing Flows, Phys. Rev. D 107(11), 113004 (2023), 10.1103/PhysRevD.107.113004, arXiv:2110.11377.
  44. Hadrons, better, faster, stronger, Mach. Learn. Sci. Tech. 3(2), 025014 (2022), 10.1088/2632-2153/ac7848, arXiv:2112.09709.
  45. Analysis-Specific Fast Simulation at the LHC with Deep Learning, Comput. Softw. Big Sci. 5(1), 15 (2021), 10.1007/s41781-021-00060-4.
  46. A. Adelmann et al., New directions for surrogate models and differentiable programming for High Energy Physics detector simulation, In Snowmass 2021 (2022), arXiv:2203.08806.
  47. V. Mikuni and B. Nachman, Score-based generative models for calorimeter shower simulation, Phys. Rev. D 106(9), 092009 (2022), 10.1103/PhysRevD.106.092009, arXiv:2206.11898.
  48. Deep generative models for fast photon shower simulation in ATLAS (2022), arXiv:2210.06204.
  49. C. Krause, I. Pang and D. Shih, CaloFlow for CaloChallenge Dataset 1 (2022), arXiv:2210.14245.
  50. CaloMan: Fast generation of calorimeter showers with density estimation on learned manifolds, In 36th Conference on Neural Information Processing Systems (2022), arXiv:2211.15380.
  51. L2LFlows: generating high-fidelity 3D calorimeter images, JINST 18(10), P10017 (2023), 10.1088/1748-0221/18/10/P10017, arXiv:2302.11594.
  52. Ultra-High-Resolution Detector Simulation with Intra-Event Aware GAN and Self-Supervised Relational Reasoning (2023), arXiv:2303.08046.
  53. Generative Machine Learning for Detector Response Modeling with a Conditional Normalizing Flow (2023), arXiv:2303.10148.
  54. New Angles on Fast Calorimeter Shower Simulation (2023), arXiv:2303.18150.
  55. CaloClouds: Fast Geometry-Independent Highly-Granular Calorimeter Simulation (2023), arXiv:2305.04847.
  56. Inductive CaloFlow (2023), arXiv:2305.11934.
  57. V. Mikuni and B. Nachman, CaloScore v2: Single-shot Calorimeter Shower Simulation with Diffusion Models (2023), arXiv:2308.03847.
  58. O. Amram and K. Pedro, Denoising diffusion models with geometry adaptation for high fidelity calorimeter simulation, Phys. Rev. D 108(7), 072014 (2023), 10.1103/PhysRevD.108.072014, arXiv:2308.03876.
  59. S. Diefenbacher, V. Mikuni and B. Nachman, Refining Fast Calorimeter Simulations with a Schrödinger Bridge (2023), arXiv:2308.12339.
  60. M. Faucci Giannelli and R. Zhang, CaloShowerGAN, a Generative Adversarial Networks model for fast calorimeter shower simulation (2023), arXiv:2309.06515.
  61. SuperCalo: Calorimeter shower super-resolution (2023), arXiv:2308.11700.
  62. Event Generation and Statistical Sampling for Physics with Deep Generative Models and a Density Information Buffer, Nature Commun. 12(1), 2985 (2021), 10.1038/s41467-021-22616-z, arXiv:1901.00875.
  63. LHC analysis-specific datasets with Generative Adversarial Networks (2019), arXiv:1901.05282.
  64. DijetGAN: A Generative-Adversarial Network Approach for the Simulation of QCD Dijet Events at the LHC, JHEP 08, 110 (2019), 10.1007/JHEP08(2019)110, arXiv:1903.02433.
  65. A. Butter, T. Plehn and R. Winterhalder, How to GAN LHC Events, SciPost Phys. 7(6), 075 (2019), 10.21468/SciPostPhys.7.6.075, arXiv:1907.03764.
  66. Simulation of electron-proton scattering events by a Feature-Augmented and Transformed Generative Adversarial Network (FAT-GAN) (2020), arXiv:2001.11103.
  67. Generative Networks for Precision Enthusiasts (2021), arXiv:2110.13632.
  68. Jet Diffusion versus JetGPT – Modern Networks for the LHC (2023), arXiv:2305.10475.
  69. DCTRGAN: Improving the Precision of Generative Models with Reweighting, JINST 15(11), P11004 (2020), 10.1088/1748-0221/15/11/P11004, arXiv:2009.03796.
  70. R. Winterhalder, M. Bellagente and B. Nachman, Latent Space Refinement for Deep Generative Models (2021), arXiv:2106.00792.
  71. B. Nachman and R. Winterhalder, ELSA – Enhanced latent spaces for improved collider simulations (2023), arXiv:2305.07696.
  72. PC-Droid: Faster diffusion and improved quality for particle cloud generation (2023), arXiv:2307.06836.
  73. How to Understand Limitations of Generative Networks (2023), arXiv:2305.16774.
  74. Measuring QCD Splittings with Invertible Networks, SciPost Phys. 10(6), 126 (2021), 10.21468/SciPostPhys.10.6.126, arXiv:2012.09873.
  75. Two Invertible Networks for the Matrix Element Method, SciPost Phys. 15, 094 (2023), 10.21468/SciPostPhys.15.3.094, arXiv:2210.00019.
  76. Precision-Machine Learning for the Matrix Element Method (2023), arXiv:2310.07752.
  77. K. Datta, D. Kar and D. Roy, Unfolding with Generative Adversarial Networks (2018), arXiv:1806.00433.
  78. How to GAN away Detector Effects, SciPost Phys. 8(4), 070 (2020), 10.21468/SciPostPhys.8.4.070, arXiv:1912.00477.
  79. OmniFold: A Method to Simultaneously Unfold All Observables, Phys. Rev. Lett. 124(18), 182001 (2020), 10.1103/PhysRevLett.124.182001, arXiv:1911.09107.
  80. Invertible Networks or Partons to Detector and Back Again, SciPost Phys. 9, 074 (2020), 10.21468/SciPostPhys.9.5.074, arXiv:2006.06685.
  81. An unfolding method based on conditional Invertible Neural Networks (cINN) using iterative training (2022), arXiv:2212.08674.
  82. ν𝜈\nuitalic_ν-flows: Conditional neutrino regression, SciPost Phys. 14(6), 159 (2023), 10.21468/SciPostPhys.14.6.159, arXiv:2207.00664.
  83. ν2superscript𝜈2\nu^{2}italic_ν start_POSTSUPERSCRIPT 2 end_POSTSUPERSCRIPT-Flows: Fast and improved neutrino reconstruction in multi-neutrino final states with conditional normalizing flows (2023), arXiv:2307.02405.
  84. End-To-End Latent Variational Diffusion Models for Inverse Problems in High Energy Physics (2023), arXiv:2305.10399.
  85. Returning CP-Observables to The Frames They Belong (2023), arXiv:2308.00027.
  86. Improving Generative Model-based Unfolding with Schrödinger Bridges (2023), arXiv:2308.12351.
  87. B. Nachman and D. Shih, Anomaly Detection with Density Estimation, Phys. Rev. D 101, 075042 (2020), 10.1103/PhysRevD.101.075042, arXiv:2001.04990.
  88. Classifying anomalies through outer density estimation, Phys. Rev. D 106(5), 055006 (2022), 10.1103/PhysRevD.106.055006, arXiv:2109.00546.
  89. CURTAINs for your sliding window: Constructing unobserved regions by transforming adjacent intervals, Front. Big Data 6, 899345 (2023), 10.3389/fdata.2023.899345, arXiv:2203.09470.
  90. Resonant anomaly detection without background sculpting, Phys. Rev. D 107(11), 114012 (2023), 10.1103/PhysRevD.107.114012, arXiv:2210.14924.
  91. FETA: Flow-Enhanced Transportation for Anomaly Detection (2022), arXiv:2212.11285.
  92. CURTAINs Flows For Flows: Constructing Unobserved Regions with Maximum Likelihood Estimation (2023), arXiv:2305.04646.
  93. Kicking it Off(-shell) with Direct Diffusion (2023), arXiv:2311.17175.
  94. CaloClouds II: Ultra-Fast Geometry-Independent Highly-Granular Calorimeter Simulation (2023), arXiv:2309.05704.
  95. Flow Matching Beyond Kinematics: Generating Jets with Particle-ID and Trajectory Displacement Information (2023), arXiv:2312.00123.
  96. Fast calorimeter simulation challenge 2022 - dataset 1, https://doi.org/10.5281/zenodo.6234054 (2022).
  97. Fast calorimeter simulation challenge 2022 - dataset 1 version 3, https://doi.org/10.5281/zenodo.8099322 (2023).
  98. Fast calorimeter simulation challenge 2022 - dataset 2, https://doi.org/10.5281/zenodo.6366271 (2022).
  99. Fast calorimeter simulation challenge 2022 - dataset 3, https://doi.org/10.5281/zenodo.6366324 (2022).
  100. Fast calorimeter simulation challenge 2022 github page, https://github.com/CaloChallenge/homepage (2022).
  101. L. Dinh, D. Krueger and Y. Bengio, NICE: Non-linear Independent Components Estimation, arXiv e-prints arXiv:1410.8516 (2014), arXiv:1410.8516.
  102. L. Dinh, J. Sohl-Dickstein and S. Bengio, Density estimation using Real NVP, arXiv e-prints arXiv:1605.08803 (2016), arXiv:1605.08803.
  103. D. P. Kingma and P. Dhariwal, Glow: Generative flow with invertible 1x1 convolutions (2018), arXiv:1807.03039.
  104. Analyzing inverse problems with invertible neural networks (2018), arXiv:1808.04730.
  105. Guided image generation with conditional invertible neural networks (2019), arXiv:1907.02392.
  106. P. Esser, R. Rombach and B. Ommer, A disentangling invertible interpretation network for explaining latent representations (2020), arXiv:2004.13166.
  107. M. F. Giannelli, private communication (2023).
  108. Neural spline flows, Advances in Neural Information Processing Systems 32, 7511 (2019), 1906.04032.
  109. Cubic-spline flows, arXiv preprint arXiv:1906.02145 (2019), 1906.02145.
  110. Framework for Easily Invertible Architectures (FrEIA) (2018-2022).
  111. G. Loaiza-Ganem and J. P. Cunningham, The continuous bernoulli: fixing a pervasive error in variational autoencoders (2019), arXiv:1907.06845.
  112. G. M. Kurtzer, V. Sochat and M. W. Bauer, Singularity: Scientific containers for mobility of compute, PLOS ONE 12(5), 1 (2017), 10.1371/journal.pone.0177459, https://doi.org/10.1371/journal.pone.0177459.
  113. Understanding Event-Generation Networks via Uncertainties, SciPost Phys. 13(1), 003 (2022), 10.21468/SciPostPhys.13.1.003, arXiv:2104.04543.
  114. Do bayesian neural networks need to be fully stochastic? (2023), arXiv:2211.06291.
Citations (15)
View on Semantic Scholar

Summary

We haven't generated a summary for this paper yet.

Ai Sparkles Streamline Icon: https://streamlinehq.com Summarize Now
X Twitter Logo Streamline Icon: https://streamlinehq.com

Tweets