Shared Data and Algorithms for Deep Learning in Fundamental Physics

We introduce a collection of datasets from fundamental physics research — including particle physics, astroparticle physics, and hadronand nuclear physics — for supervised machine learning studies. These datasets, containing hadronic top quarks, cosmic-ray induced air showers, phase transitions in hadronic matter, and generator-level histories, are made public to simplify future work on cross-disciplinary machine learning and transfer learning in fundamental physics. Based on these data, we present a simple yet flexible graph-based neural network architecture that can easily be applied to a wide range of supervised learning tasks in these domains. We show that our approach reaches performance close to state-of-the-art dedicated methods on all datasets. To simplify adaptation for various problems, we provide easy-to-follow instructions on how graph-based representations of data structures, relevant for fundamental physics, can be constructed and provide code implementations for several of them. Implementations are also provided for our proposed method and all reference algorithms. L. Benato, E. Buhmann, G. Kasieczka, and W. Korcari Institut für Experimentalphysik, Universität Hamburg, Germany E-mail: gregor.kasieczka@uni-hamburg.de J. Glombitza, M. Erdmann, and P. Fackeldey III. Physikalisches Institut A, RWTH Aachen University, Germany N. Hartmann and T. Kuhr Fakultät für Physik, Ludwig Maximilians University Munich, Germany J. Steinheimer, H. Stöcker, and K. Zhou Frankfurt Institute for Advanced Studies (FIAS), Germany T. Plehn Institut für Theoretische Physik, Universität Heidelberg, Germany

[1]  R. Bellwied,et al.  Observables and predictions , 2011 .

[2]  J. Erdmann,et al.  A tagger for strange jets based on tracking information using long short-term memory , 2019, Journal of Instrumentation.

[3]  Charanjit K. Khosa,et al.  The LHC Olympics 2020 a community challenge for anomaly detection in high energy physics , 2021, Reports on progress in physics. Physical Society.

[4]  David W. Miller,et al.  Lorentz Group Equivariant Neural Network for Particle Physics , 2020, ICML.

[5]  Balázs Kégl,et al.  The Higgs boson machine learning challenge , 2014, HEPML@NIPS.

[6]  Max Welling,et al.  Semi-Supervised Classification with Graph Convolutional Networks , 2016, ICLR.

[7]  M. Erdmann,et al.  Lorentz Boost Networks: autonomous physics-inspired feature engineering , 2018, Journal of Instrumentation.

[8]  Lijia Jiang,et al.  Deep learning stochastic processes with QCD phase transition , 2021, Physical Review D.

[9]  Luke de Oliveira,et al.  Jet-images — deep learning edition , 2015, Journal of High Energy Physics.

[10]  G. Kasieczka,et al.  Deep-learning top taggers or the end of QCD? , 2017, 1701.08784.

[11]  Gregor Kasieczka,et al.  Deep-learned Top Tagging with a Lorentz Layer , 2017, SciPost Physics.

[12]  David Rousseau,et al.  Machine Learning Scientific Competitions and Datasets , 2020, Artificial Intelligence for High Energy Physics.

[13]  Cécile Germain,et al.  The Tracking Machine Learning Challenge: Accuracy Phase , 2019, The NeurIPS '18 Competition.

[14]  Huilin Qu,et al.  ParticleNet: Jet Tagging via Particle Clouds , 2019, Physical Review D.

[15]  Yue Wang,et al.  Dynamic Graph CNN for Learning on Point Clouds , 2018, ACM Trans. Graph..

[16]  H. Stöcker,et al.  An equation-of-state-meter of quantum chromodynamics transition from deep learning , 2018, Nature Communications.

[17]  H. Stoecker,et al.  Identifying the nature of the QCD transition in heavy-ion collisions with deep learning , 2020, Nuclear Physics A.

[19]  Andrew Zisserman,et al.  Very Deep Convolutional Networks for Large-Scale Image Recognition , 2014, ICLR.

[20]  M. Erdmann,et al.  A deep learning-based reconstruction of cosmic ray-induced air showers , 2017, 1708.00647.

[21]  M. Cacciari,et al.  The anti-$k_t$ jet clustering algorithm , 2008, 0802.1189.

[22]  Selective background Monte Carlo simulation at Belle II , 2020 .

[23]  Patrick T. Komiske,et al.  Energy flow networks: deep sets for particle jets , 2018, Journal of High Energy Physics.

[24]  J. Randrup,et al.  Spinodal amplification of density fluctuations in fluid-dynamical simulations of relativistic nuclear collisions. , 2012, Physical review letters.

[25]  H. Stoecker,et al.  A machine learning study on spinodal clumping in heavy ion collisions , 2021 .

[26]  H. Stöcker,et al.  Identifying the nature of the QCD transition in relativistic collision of heavy nuclei with deep learning , 2019, The European Physical Journal C.

[27]  M. Pierini,et al.  The Dark Machines Anomaly Score Challenge: Benchmark Data and Model Independent Event Classification for the Large Hadron Collider , 2021, SciPost Physics.

[28]  T. Lohse,et al.  Application of deep learning methods to analysis of imaging atmospheric Cherenkov telescopes data , 2018, Astroparticle Physics.

[29]  M. Cacciari,et al.  FastJet user manual , 2011, 1111.6097.

[31]  Hayes,et al.  Review of Particle Physics. , 1996, Physical review. D, Particles and fields.

[32]  Prabhat,et al.  Graph Neural Networks for Particle Reconstruction in High Energy Physics detectors , 2020, 2003.11603.

[33]  H. Stöcker,et al.  Classify QCD phase transition with deep learning , 2019, Nuclear Physics A.

[34]  Shannon Egan,et al.  Long Short-Term Memory (LSTM) networks with jet constituents for boosted top tagging at the LHC , 2017, ArXiv.

[35]  Leandro Giordano Almeida,et al.  Playing tag with ANN: boosted top identification with pattern recognition , 2015, 1501.05968.

[36]  James Meier Samuel Kahn,et al.  Selective background Monte Carlo simulation at Belle II , 2019, Journal of Physics: Conference Series.

[37]  Jimmy Ba,et al.  Adam: A Method for Stochastic Optimization , 2014, ICLR.

[38]  Jonathan Shlomi,et al.  Graph neural networks in particle physics , 2020, Mach. Learn. Sci. Technol..

[39]  L. Pang,et al.  Regressive and generative neural networks for scalar field theory , 2018, Physical Review D.

[40]  Maria Spiropulu,et al.  Pileup mitigation at the Large Hadron Collider with graph neural networks , 2018, The European Physical Journal Plus.

[41]  Alex Krizhevsky,et al.  Learning Multiple Layers of Features from Tiny Images , 2009 .

[43]  Javier Duarte,et al.  Graph Neural Networks for Particle Tracking and Reconstruction , 2020 .

[44]  Peter Skands,et al.  An introduction to PYTHIA 8.2 , 2014, Comput. Phys. Commun..

[45]  L. Gouskos,et al.  The Machine Learning landscape of top taggers , 2019, SciPost Physics.

[46]  Prabhat,et al.  Track Seeding and Labelling with Embedded-space Graph Neural Networks , 2020, ArXiv.

[47]  Li Fei-Fei,et al.  ImageNet: A large-scale hierarchical image database , 2009, CVPR.

[48]  Nitish Srivastava,et al.  Dropout: a simple way to prevent neural networks from overfitting , 2014, J. Mach. Learn. Res..

[49]  Yutaro Iiyama,et al.  Learning representations of irregular particle-detector geometry with distance-weighted graph networks , 2019, The European Physical Journal C.

[50]  L. Gouskos,et al.  Jet Tagging via Particle Clouds , 2021 .

[51]  Javier Mauricio Duarte,et al.  Graph Generative Adversarial Networks for Sparse Data Generation in High Energy Physics , 2020, ArXiv.

[52]  Jian Sun,et al.  Delving Deep into Rectifiers: Surpassing Human-Level Performance on ImageNet Classification , 2015, 2015 IEEE International Conference on Computer Vision (ICCV).

[53]  Patrick T. Komiske,et al.  Deep learning in color: towards automated quark/gluon jet discrimination , 2016, Journal of High Energy Physics.

[54]  Carlos Escobar.  THE PIERRE AUGER COSMIC RAY OBSERVATORY , 1998 .

[55]  F. Kahlhoefer,et al.  Casting a graph net to catch dark showers , 2020, SciPost Physics.

[56]  D. Lange,et al.  The EvtGen particle decay simulation package , 2001 .

[57]  Eric A. Moreno,et al.  JEDI-net: a jet identification algorithm based on interaction networks , 2019, The European Physical Journal C.

[58]  Volker Koch,et al.  A machine learning study to identify spinodal clumping in high energy nuclear collisions , 2019, Journal of High Energy Physics.

[59]  P. G. Isar,et al.  Deep-learning based reconstruction of the shower maximum X max using the water-Cherenkov detectors of the Pierre Auger Observatory , 2021, Journal of Instrumentation.

[60]  J. Favereau,et al.  DELPHES 3: a modular framework for fast simulation of a generic collider experiment , 2013, Journal of High Energy Physics.

[61]  Javier Duarte,et al.  Accelerated Charged Particle Tracking with Graph Neural Networks on FPGAs , 2020, ArXiv.

[62]  Luke de Oliveira,et al.  Electromagnetic Calorimeter Shower Images , 2017 .

[63]  Maria Spiropulu,et al.  MLPF: efficient machine-learned particle-flow reconstruction using graph neural networks , 2021, The European Physical Journal C.