SciPost Physics

Deep-learned Top Tagging with a Lorentz Layer

Anja Butter, Gregor Kasieczka, Tilman Plehn, Michael Russell

SciPost Phys. 5, 028 (2018) · published 26 September 2018

• doi: 10.21468/SciPostPhys.5.3.028
• pdf
• BiBTeX
• RIS
• Submissions/Reports
• 

Abstract

We introduce a new and highly efficient tagger for hadronically decaying top quarks, based on a deep neural network working with Lorentz vectors and the Minkowski metric. With its novel machine learning setup and architecture it allows us to identify boosted top quarks not only from calorimeter towers, but also including tracking information. We show how the performance of our tagger compares with QCD-inspired and image-recognition approaches and find that it significantly increases the performance for strongly boosted top quarks.

• Lin et al., Boosting $$ H\to b\overline{b} $$ with machine learning
  J. High Energ. Phys. 2018, 101 (2018) [Crossref]
• Andreassen et al., binary junipr: An Interpretable Probabilistic Model for Discrimination
  Phys. Rev. Lett. 123, 182001 (2019) [Crossref]
• Kasieczka et al., Automating the ABCD method with machine learning
  Phys. Rev. D 103, 035021 (2021) [Crossref]
• Macaluso et al., Pulling out all the tops with computer vision and deep learning
  J. High Energ. Phys. 2018, 121 (2018) [Crossref]
• Atkinson et al., IRC-Safe Graph Autoencoder for Unsupervised Anomaly Detection
  Front. Artif. Intell. 5, 943135 (2022) [Crossref]
• Komiske et al., An operational definition of quark and gluon jets
  J. High Energ. Phys. 2018, 59 (2018) [Crossref]
• Collins et al., Extending the search for new resonances with machine learning
  Phys. Rev. D 99, 014038 (2019) [Crossref]
• Abdughani et al., Supervised Deep Learning in High Energy Phenomenology: a Mini Review*
  Commun. Theor. Phys. 71, 955 (2019) [Crossref]
• Piscopo et al., Solving differential equations with neural networks: Applications to the calculation of cosmological phase transitions
  Phys. Rev. D 100, 016002 (2019) [Crossref]
• Ju et al., Supervised jet clustering with graph neural networks for Lorentz boosted bosons
  Phys. Rev. D 102, 075014 (2020) [Crossref]
• Konar et al., Energy-weighted message passing: an infra-red and collinear safe graph neural network algorithm
  J. High Energ. Phys. 2022, 60 (2022) [Crossref]
• Han et al., Six top messages of new physics at the LHC
  J. High Energ. Phys. 2019, 8 (2019) [Crossref]
• Tan et al., Machine learning phases and criticalities without using real data for training
  Phys. Rev. B 102, 224434 (2020) [Crossref]
• Bernreuther et al., Casting a graph net to catch dark showers
  SciPost Phys. 10, 046 (2021) [Crossref]
• Heimel et al., QCD or what?
  SciPost Phys. 6, 030 (2019) [Crossref]
• Datta et al., Automating the construction of jet observables with machine learning
  Phys. Rev. D 100, 095016 (2019) [Crossref]
• Chakraborty et al., Interpretable deep learning for two-prong jet classification with jet spectra
  J. High Energ. Phys. 2019, 135 (2019) [Crossref]
• Lester et al., Lorentz and permutation invariants of particles III: Constraining nonstandard sources of parity violation
  Int. J. Mod. Phys. A 37, 2250093 (2022) [Crossref]
• Kasieczka et al., Quark-gluon tagging: Machine learning vs detector
  SciPost Phys. 6, 069 (2019) [Crossref]
• D’Agnolo et al., Learning new physics from a machine
  Phys. Rev. D 99, 015014 (2019) [Crossref]
• Shlomi et al., Graph neural networks in particle physics
  Mach. Learn.: Sci. Technol. 2, 021001 (2021) [Crossref]
• Araz et al., Combine and conquer: event reconstruction with Bayesian Ensemble Neural Networks
  J. High Energ. Phys. 2021, 296 (2021) [Crossref]
• Pol et al., Lightweight jet reconstruction and identification as an object detection task
  Mach. Learn.: Sci. Technol. 3, 025016 (2022) [Crossref]
• Amram et al., Tag N’ Train: a technique to train improved classifiers on unlabeled data
  J. High Energ. Phys. 2021, 153 (2021) [Crossref]
• Dreyer et al., Jet tagging in the Lund plane with graph networks
  J. High Energ. Phys. 2021, 52 (2021) [Crossref]
• Dasgupta et al., Investigating top tagging with Ym-Splitter and N-subjettiness
  J. High Energ. Phys. 2021, 92 (2021) [Crossref]
• Andrews,
  , 59 (2023) [Crossref]
• Kar et al., Exploring jet substructure in semi-visible jets
  SciPost Phys. 10, 084 (2021) [Crossref]
• Benato et al., Teaching machine learning with an application in collider particle physics
  J. Inst. 15, C09011 (2020) [Crossref]
• Bieringer et al., Measuring QCD Splittings with Invertible Networks
  SciPost Phys. 10, 126 (2021) [Crossref]
• Hajer et al., Novelty detection meets collider physics
  Phys. Rev. D 101, 076015 (2020) [Crossref]
• Nguyen et al., Topology Classification with Deep Learning to Improve Real-Time Event Selection at the LHC
  Comput Softw Big Sci 3, 12 (2019) [Crossref]
• Chen et al., Boosted W and Z tagging with jet charge and deep learning
  Phys. Rev. D 101, 053001 (2020) [Crossref]
• Nachman, A guide for deploying Deep Learning in LHC searches: How to achieve optimality and account for uncertainty
  SciPost Phys. 8, 090 (2020) [Crossref]
• Kasieczka et al., The Machine Learning landscape of top taggers
  SciPost Phys. 7, 014 (2019) [Crossref]
• Duarte et al., FPGA-Accelerated Machine Learning Inference as a Service for Particle Physics Computing
  Comput Softw Big Sci 3, 13 (2019) [Crossref]
• Louppe et al., QCD-aware recursive neural networks for jet physics
  J. High Energ. Phys. 2019, 57 (2019) [Crossref]
• Li et al., Tag Z boson jets via convolutional neural networks
  Acta Phys. Sin. 70, 061301 (2021) [Crossref]
• Larkoski et al., Jet substructure at the Large Hadron Collider: A review of recent advances in theory and machine learning
  Physics Reports 841, 1 (2020) [Crossref]
• Biscarat et al., Jet Single Shot Detection
  EPJ Web Conf. 251, 04027 (2021) [Crossref]
• Badea et al., Solving combinatorial problems at particle colliders using machine learning
  Phys. Rev. D 106, 016001 (2022) [Crossref]
• Aaboud et al., Performance of top-quark and $$\varvec{W}$$ W -boson tagging with ATLAS in Run 2 of the LHC
  Eur. Phys. J. C 79, 375 (2019) [Crossref]
• Chatterjee et al., Jets with electrons from boosted top quarks
  J. High Energ. Phys. 2020, 170 (2020) [Crossref]
• Chiang et al., Vector boson fusion versus gluon-gluon fusion Higgs boson production with full-event deep learning: Toward a decay-agnostic tagger
  Phys. Rev. D 107, 016014 (2023) [Crossref]
• Gong et al., An efficient Lorentz equivariant graph neural network for jet tagging
  J. High Energ. Phys. 2022, 30 (2022) [Crossref]
• Biscarat et al., End-to-End Jet Classification of Boosted Top Quarks with CMS Open Data
  EPJ Web Conf. 251, 04030 (2021) [Crossref]
• Moreno et al., Interaction networks for the identification of boosted H→bb¯ decays
  Phys. Rev. D 102, 012010 (2020) [Crossref]
• Dolan et al., Equivariant energy flow networks for jet tagging
  Phys. Rev. D 103, 074022 (2021) [Crossref]
• von Rueden et al., Informed Machine Learning - A Taxonomy and Survey of Integrating Prior Knowledge into Learning Systems
  IEEE Trans. Knowl. Data Eng., 1 (2021) [Crossref]
• Choi et al., Infrared safety of a neural-net top tagging algorithm
  J. High Energ. Phys. 2019, 132 (2019) [Crossref]
• Kasieczka et al., Robust Jet Classifiers through Distance Correlation
  Phys. Rev. Lett. 125, 122001 (2020) [Crossref]
• Erdmann et al., Maximum performance of strange-jet tagging at hadron colliders
  J. Inst. 16, P08039 (2021) [Crossref]
• Andreassen et al., Neural networks for full phase-space reweighting and parameter tuning
  Phys. Rev. D 101, 091901 (2020) [Crossref]
• Doglioni et al., Physics Inspired Deep Neural Networks for Top Quark Reconstruction
  EPJ Web Conf. 245, 06029 (2020) [Crossref]
• Bedeschi et al., Jet flavour tagging for future colliders with fast simulation
  Eur. Phys. J. C 82, 646 (2022) [Crossref]
• Luo et al., Quark jet versus gluon jet: fully-connected neural networks with high-level features
  Sci. China Phys. Mech. Astron. 62, 991011 (2019) [Crossref]
• Brauer et al.,
  , 1437 (2022) [Crossref]
• Chakraborty et al., Neural network-based top tagger with two-point energy correlations and geometry of soft emissions
  J. High Energ. Phys. 2020, 111 (2020) [Crossref]
• Dillon et al., Symmetries, safety, and self-supervision
  SciPost Phys. 12, 188 (2022) [Crossref]
• Moreno et al., JEDI-net: a jet identification algorithm based on interaction networks
  Eur. Phys. J. C 80, 58 (2020) [Crossref]
• Kasieczka et al., Per-object systematics using deep-learned calibration
  SciPost Phys. 9, 089 (2020) [Crossref]
• Tan et al., A universal neural network for learning phases
  Eur. Phys. J. Plus 136, 1116 (2021) [Crossref]
• Shen, A Transdisciplinary Review of Deep Learning Research and Its Relevance for Water Resources Scientists
  Water Resour. Res. 54, 8558 (2018) [Crossref]
• Andrews et al., End-to-end jet classification of boosted top quarks with the CMS open data
  Phys. Rev. D 105, 052008 (2022) [Crossref]
• Erdmann et al., Lorentz Boost Networks: autonomous physics-inspired feature engineering
  J. Inst. 14, P06006 (2019) [Crossref]
• Erdmann et al., Reinforced sorting networks for particle physics analyses
  J. Phys.: Conf. Ser. 1525, 012098 (2020) [Crossref]
• Batson et al., Topological obstructions to autoencoding
  J. High Energ. Phys. 2021, 280 (2021) [Crossref]
• Qu et al., Jet tagging via particle clouds
  Phys. Rev. D 101, 056019 (2020) [Crossref]
• Dreyer et al., Leveraging universality of jet taggers through transfer learning
  Eur. Phys. J. C 82, 564 (2022) [Crossref]
• Abdughani et al., Probing stop pair production at the LHC with graph neural networks
  J. High Energ. Phys. 2019, 55 (2019) [Crossref]
• Blance et al., Adversarially-trained autoencoders for robust unsupervised new physics searches
  J. High Energ. Phys. 2019, 47 (2019) [Crossref]
• Guo et al., Deep learning for R -parity violating supersymmetry searches at the LHC
  Phys. Rev. D 98, 076017 (2018) [Crossref]
• Bradshaw et al., Mass agnostic jet taggers
  SciPost Phys. 8, 011 (2020) [Crossref]
• Farina et al., Searching for new physics with deep autoencoders
  Phys. Rev. D 101, 075021 (2020) [Crossref]
• Baldi et al., How to GAN Higher Jet Resolution
  SciPost Phys. 13, 064 (2022) [Crossref]
• Dillon et al., Uncovering latent jet substructure
  Phys. Rev. D 100, 056002 (2019) [Crossref]
• Finke et al., Autoencoders for unsupervised anomaly detection in high energy physics
  J. High Energ. Phys. 2021, 161 (2021) [Crossref]
• Chen et al., Deep learning jet substructure from two-particle correlations
  Phys. Rev. D 101, 114025 (2020) [Crossref]
• Gripaios et al., Lorentz- and permutation-invariants of particles
  J. Phys. A: Math. Theor. 54, 155201 (2021) [Crossref]
• Moore et al., Reports of my demise are greatly exaggerated: $N$-subjettiness taggers take on jet images
  SciPost Phys. 7, 036 (2019) [Crossref]
• Monaco et al., Theory-Guided Deep Learning Algorithms: An Experimental Evaluation
  Electronics 11, 2850 (2022) [Crossref]
• Komiske et al., Energy flow networks: deep sets for particle jets
  J. High Energ. Phys. 2019, 121 (2019) [Crossref]
• Murnane et al., Semi-Equivariant GNN Architectures for Jet Tagging
  J. Phys.: Conf. Ser. 2438, 012121 (2023) [Crossref]
• Alves et al., Towards recognizing the light facet of the Higgs boson
  Mach. Learn.: Sci. Technol. 1, 045025 (2020) [Crossref]
• Ngairangbam et al., Invisible Higgs search through vector boson fusion: a deep learning approach
  Eur. Phys. J. C 80, 1055 (2020) [Crossref]
• Bothmann et al., Reweighting a parton shower using a neural network: the final-state case
  J. High Energ. Phys. 2019, 33 (2019) [Crossref]
• Monaco et al.,
  1652, 256 (2022) [Crossref]
• Park et al., Quasi anomalous knowledge: searching for new physics with embedded knowledge
  J. High Energ. Phys. 2021, 30 (2021) [Crossref]
• Benato et al., Shared Data and Algorithms for Deep Learning in Fundamental Physics
  Comput Softw Big Sci 6, 9 (2022) [Crossref]
• Albouy et al., Theory, phenomenology, and experimental avenues for dark showers: a Snowmass 2021 report
  Eur. Phys. J. C 82, 1132 (2022) [Crossref]
• Fraser et al., Jet charge and machine learning
  J. High Energ. Phys. 2018, 93 (2018) [Crossref]
• D’Agnolo et al., Learning multivariate new physics
  Eur. Phys. J. C 81, 89 (2021) [Crossref]
• Carrazza et al., Jet grooming through reinforcement learning
  Phys. Rev. D 100, 014014 (2019) [Crossref]
• Diefenbacher et al., CapsNets continuing the convolutional quest
  SciPost Phys. 8, 023 (2020) [Crossref]
• Bollweg et al., Deep-learning jets with uncertainties and more
  SciPost Phys. 8, 006 (2020) [Crossref]