Parameter Inference from Event Ensembles and the Top-Quark Mass

Abstract

One of the key tasks of any particle collider is measurement. In practice, this is often done by fitting data to a simulation, which depends on many parameters. Sometimes, when the effects of varying different parameters are highly correlated, a large ensemble of data may be needed to resolve parameter-space degeneracies. An important example is measuring the top-quark mass, where other physical and unphysical parameters in the simulation must be marginalized over when fitting the top-quark mass parameter. We compare three different methodologies for top-quark mass measurement: a classical histogram fitting procedure, similar to one commonly used in experiment optionally augmented with soft-drop jet grooming; a machine-learning method called DCTR; and a linear regression approach, either using a least-squares fit or with a dense linearly-activated neural network. Despite the fact that individual events are totally uncorrelated, we find that the linear regression methods work most effectively when we input an ensemble of events sorted by mass, rather than training them on individual events. Although all methods provide robust extraction of the top-quark mass parameter, the linear network does marginally best and is remarkably simple. For the top study, we conclude that the Monte-Carlo-based uncertainty on current extractions of the top-quark mass from LHC data can be reduced significantly (by perhaps a factor of 2) using networks trained on sorted event ensembles. More generally, machine learning from ensembles for parameter estimation has broad potential for collider physics measurements.

arXiv commentv1: 27 + 5 pages, 14 + 3 figures. v2: Matches version accepted to JHEP

References in corpus (55)

• Scikit-learn: Machine Learning in Python
• The anti-k_t jet clustering algorithm
• An Introduction to PYTHIA 8.2
• FastJet user manual
• New Generation of Parton Distributions with Uncertainties from Global QCD Analysis
• Higgs mass and vacuum stability in the Standard Model at NNLO
• Soft Drop
• On the metastability of the Standard Model vacuum
• Energy Flow Networks: Deep Sets for Particle Jets
• Jets from Massive Unstable Particles: Top-Mass Determination
• Parameterized Machine Learning for High-Energy Physics
• Non-perturbative QCD Effects and the Top Mass at the Tevatron
• OmniFold: A Method to Simultaneously Unfold All Observables
• Mining gold from implicit models to improve likelihood-free inference
• Constraining Effective Field Theories with Machine Learning
• Simulation Assisted Likelihood-free Anomaly Detection
• A Guide to Constraining Effective Field Theories with Machine Learning
• Measurement of the $t\bar{t}$ production cross-section using $e\mu$ events with $b$-tagged jets in $pp$ collisions at $\sqrt{s}=7$ and 8 TeV with the ATLAS detector
• Scale Invariant Instantons and the Complete Lifetime of the Standard Model
• Study of Monte Carlo approach to experimental uncertainty propagation with MSTW 2008 PDFs
• Approximating Likelihood Ratios with Calibrated Discriminative Classifiers
• Effects of color reconnection on $t\bar{t}$ final states at the LHC
• Top Mass Measurements from Jets and the Tevatron Top-Quark Mass
• Measurement of the $\mathrm{t}\overline{\mathrm{t}}$ production cross section, the top quark mass, and the strong coupling constant using dilepton events in pp collisions at $\sqrt{s}=$ 13 TeV
• JUNIPR: a Framework for Unsupervised Machine Learning in Particle Physics
• Measurement of the top quark mass in the $t\bar{t}\to$ lepton+jets channel from $\sqrt{s}=8$ TeV ATLAS data and combination with previous results
• Neural Networks for Full Phase-space Reweighting and Parameter Tuning
• Measurement of $\mathrm{t}\bar{\mathrm{t}}$ normalised multi-differential cross sections in pp collisions at $\sqrt{s}=13$ TeV, and simultaneous determination of the strong coupling strength, top quark pole mass, and parton distribution functions
• Measurement of the t-tbar production cross section in the e-mu channel in proton-proton collisions at sqrt(s) = 7 and 8 TeV
• Measurement of the top quark mass in the $t\bar{t}\to$ dilepton channel from $\sqrt{s}=8$ TeV ATLAS data
• Top Quark Mass Calibration for Monte Carlo Event Generators
• Measurement of the top quark mass in the all-jets final state at $\sqrt{s}=$ 13 TeV and combination with the lepton+jets channel
• What is the Top Quark Mass?
• Qjets: A Non-Deterministic Approach to Tree-Based Jet Substructure
• Closing the stop gap
• Measurements of top-quark pair spin correlations in the $e\mu$ channel at $\sqrt{s}=13$ TeV using $pp$ collisions in the ATLAS detector
• Extracting a Short Distance Top Mass with Light Grooming
• Measurement of the top quark mass with lepton+jets final states using pp collisions at $\sqrt{s}=$ 13 TeV
• Top-quark mass measurement in the all-hadronic $t\bar{t}$ decay channel at $\sqrt{s}$ = 8 TeV with the ATLAS detector
• On the Cutoff Dependence of the Quark Mass Parameter in Angular Ordered Parton Showers
• Permutationless Many-Jet Event Reconstruction with Symmetry Preserving Attention Networks
• Measurement of lepton differential distributions and the top quark mass in $t\bar{t}$ production in $pp$ collisions at $\sqrt{s}=8$ TeV with the ATLAS detector
• Binary JUNIPR: an interpretable probabilistic model for discrimination
• Calibration of the Top-Quark Monte-Carlo Mass
• Sneaky light stop
• Likelihood-free inference with an improved cross-entropy estimator
• Telescoping Jets: Multiple Event Interpretations with Multiple R's
• Parameter Estimation using Neural Networks in the Presence of Detector Effects
• Resonance Searches with Machine Learned Likelihood Ratios
• Reducing the Top Quark Mass Uncertainty with Jet Grooming
• On the ATLAS Top Mass Measurements and the Potential for Stealth Stop Contamination
• Magnifying the ATLAS Stealth Stop Splinter: Impact of Spin Correlations and Finite Widths
• Efficiency Parameterization with Neural Networks
• Quantifying the power of multiple event interpretations
• Parton Distributions and Event Generators

Cited by in corpus (6)

• E Pluribus Unum Ex Machina: Learning from Many Collider Events at Once
• A method for approximating optimal statistical significances with machine-learned likelihoods
• Bayesian Probabilistic Modelling for Four-Tops at the LHC
• A $W^{\pm }$ polarization analyzer from Deep Neural Networks
• BitHEP -- The Limits of Low-Precision ML in HEP
• Mixture Density Network Estimation of Continuous Variable Maximum Likelihood Using Discrete Training Samples