A guide for deploying Deep Learning in LHC searches: How to achieve optimality and account for uncertainty

Abstract

Deep learning tools can incorporate all of the available information into a search for new particles, thus making the best use of the available data. This paper reviews how to optimally integrate information with deep learning and explicitly describes the corresponding sources of uncertainty. Simple illustrative examples show how these concepts can be applied in practice.

22 pages, 7 figures. v2: expanded discussion on removing sensitivity to theory nuisance parameters. v3: Updated with suggestions from referees

References in corpus (38)

• Observation of a new particle in the search for the Standard Model Higgs boson with the ATLAS detector at the LHC
• Observation of a new boson at a mass of 125 GeV with the CMS experiment at the LHC
• Jet energy scale and resolution in the CMS experiment in pp collisions at 8 TeV
• Jet energy measurement and its systematic uncertainty in proton-proton collisions at $\sqrt{s}=7$ TeV with the ATLAS detector
• Anomaly Detection with Density Estimation
• Classification without labels: Learning from mixed samples in high energy physics
• Extending the Bump Hunt with Machine Learning
• OmniFold: A Method to Simultaneously Unfold All Observables
• A generic anti-QCD jet tagger
• Simulation Assisted Likelihood-free Anomaly Detection
• Parton Shower Uncertainties in Jet Substructure Analyses with Deep Neural Networks
• JEDI-net: a jet identification algorithm based on interaction networks
• Pileup Mitigation with Machine Learning (PUMML)
• Tag N' Train: A Technique to Train Improved Classifiers on Unlabeled Data
• Neural Networks for Full Phase-space Reweighting and Parameter Tuning
• How to GAN away Detector Effects
• Reweighting with Boosted Decision Trees
• DisCo Fever: Robust Networks Through Distance Correlation
• Interaction networks for the identification of boosted $H\to b\overline{b}$ decays
• Fast simulation of muons produced at the SHiP experiment using Generative Adversarial Networks
• Lund jet images from generative and cycle-consistent adversarial networks
• Boosted $W/Z$ Tagging with Jet Charge and Deep Learning
• Lorentz Boost Networks: Autonomous Physics-Inspired Feature Engineering
• Power-Constrained Limits
• Mass Agnostic Jet Taggers
• Search for non-resonant Higgs boson pair production in the $bb\ell \nu \ell \nu$ final state with the ATLAS detector in $pp$ collisions at $\sqrt{s}=13$ TeV
• LHC analysis-specific datasets with Generative Adversarial Networks
• Long Short-Term Memory (LSTM) networks with jet constituents for boosted top tagging at the LHC
• Cherenkov Detectors Fast Simulation Using Neural Networks
• The DNNLikelihood: enhancing likelihood distribution with Deep Learning
• Quark-Gluon Jet Discrimination Using Convolutional Neural Networks
• Machine Learning Templates for QCD Factorization in the Search for Physics Beyond the Standard Model
• Jet grooming through reinforcement learning
• Learning Multivariate New Physics
• Machine learning approach to inverse problem and unfolding procedure
• Investigating the Quantum Properties of Jets and the Search for a Supersymmetric Top Quark Partner with the ATLAS Detector
• Towards automation of data quality system for CERN CMS experiment
• Importance and construction of features in identifying new physics signals with deep learning

Cited by in corpus (29)

• OmniFold: A Method to Simultaneously Unfold All Observables
• Simulation Assisted Likelihood-free Anomaly Detection
• Machine Learning and LHC Event Generation
• Understanding Event-Generation Networks via Uncertainties
• Generative Networks for Precision Enthusiasts
• A Neural Resampler for Monte Carlo Reweighting with Preserved Uncertainties
• The Present and Future of QCD
• A factorisation-aware Matrix element emulator
• Casting a graph net to catch dark showers
• Uncertainty Aware Learning for High Energy Physics
• Learning the latent structure of collider events
• Mapping Machine-Learned Physics into a Human-Readable Space
• Presenting Unbinned Differential Cross Section Results
• Per-Object Systematics using Deep-Learned Calibration
• The LHC Olympics 2020: A Community Challenge for Anomaly Detection in High Energy Physics
• Boosted decision trees in the era of new physics: a smuon analysis case study
• Energy-weighted Message Passing: an infra-red and collinear safe graph neural network algorithm
• Towards Machine Learning Analytics for Jet Substructure
• A Cautionary Tale of Decorrelating Theory Uncertainties
• Combine and Conquer: Event Reconstruction with Bayesian Ensemble Neural Networks
• Learning Likelihood Ratios with Neural Network Classifiers
• Null Hypothesis Test for Anomaly Detection
• A method for approximating optimal statistical significances with machine-learned likelihoods
• Bias and Priors in Machine Learning Calibrations for High Energy Physics
• Interpretable deep learning models for the inference and classification of LHC data
• Learning Feynman integrals from differential equations with neural networks
• Exploring the Universality of Hadronic Jet Classification
• Observable Optimization for Precision Theory: Machine Learning Energy Correlators
• Boosted and semi-boosted all-hadronic $t\bar{t}$ reconstruction performance on kinematic variables for selected BSM models using a 2D extesion of the BumpHunter algorithm