Neural simulation-based inference (NSBI) is a powerful class of machine-learning-based methods for statistical inference that naturally handles high-dimensional parameter estimation without the need to bin data into low-dimensional summary histograms. Such methods are promising for a range of measurements, including at the Large Hadron Collider, where no single observable may be optimal to scan over the entire theoretical phase space under consideration, or where binning data into histograms could result in a loss of sensitivity. This work develops a NSBI framework for statistical inference, using neural networks to estimate probability density ratios, which enables the application to a full-scale analysis. It incorporates a large number of systematic uncertainties, quantifies the uncertainty due to the finite number of events in training samples, develops a method to construct confidence intervals, and demonstrates a series of intermediate diagnostic checks that can be performed to validate the robustness of the method. As an example, the power and feasibility of the method are assessed on simulated data for a simplified version of an off-shell Higgs boson couplings measurement in the four-lepton final states. This approach represents an extension to the standard statistical methodology used by the experiments at the Large Hadron Collider, and can benefit many physics analyses.

This report reviews the published results of searches for possible additional scalar particles and exotic decays of the Higgs boson performed by the ATLAS Collaboration using up to 140 fb⁻¹ of 13 TeV proton–proton collision data collected during Run 2 of the Large Hadron Collider. Key results are examined, and observed excesses, while never statistically compelling, are noted. Constraints are placed on parameters of several models which extend the Standard Model, for example by adding one or more singlet or doublet fields, or offering exotic Higgs boson decay channels. Summaries of new searches as well as extensions of previous searches are discussed. These new results have a wider reach or attain stronger exclusion limits. New experimental techniques that were developed for these searches are highlighted. Search channels which have not yet been examined are also listed, as these provide insight into possible future areas of exploration.

A combination of searches for the single production of vectorlike top quarks (𝑇) is presented. These analyses are based on proton-proton collisions at √𝑠 =13  TeV recorded in 2015–2018 with the ATLAS detector at the Large Hadron Collider, corresponding to an integrated luminosity of 139  fb⁻¹. The 𝑇 decay modes considered in this combination are into a top quark and either a Standard Model Higgs boson or a 𝑍 boson (𝑇 →𝐻⁡𝑡 and 𝑇→𝑍⁢𝑡). The individual searches used in the combination are differentiated by the number of leptons (𝑒, 𝜇) in the final state. The observed data are found to be in good agreement with the Standard Model background prediction. Interpretations are provided for a range of masses and couplings of the vectorlike top quark for benchmark models and generalized representations in terms of 95% confidence level limits. For a benchmark signal prediction of a vectorlike top quark SU(2) singlet with electroweak coupling, 𝜅, of 0.5, masses below 2.1 TeV are excluded, resulting in the most restrictive limits to date.

The production of D^± and D^±s charmed mesons is measured using the D^±/D^±s → ϕ(μμ)π^± decay channel with 137 fb⁻¹ of √s = 13 TeV proton-proton collision data collected with the ATLAS detector at the Large Hadron Collider during the years 2016–2018. The charmed mesons are reconstructed in the range of transverse momentum 12 < p_T < 100 GeV and pseudorapidity |η| < 2.5. The differential cross-sections are measured as a function of transverse momentum and pseudorapidity, respectively, and compared with next-to-leading- order QCD predictions. The predictions are found to be consistent with the measurements in the visible kinematic region within the large theoretical uncertainties.

This paper presents a new τ-lepton reconstruction and identification procedure at the ATLAS detector at the Large Hadron Collider, which leads to significantly improved performance in the case of physics processes where a highly boosted pair of τ-leptons is produced and one τ-lepton decays into a muon and two neutrinos (τ_μ), and the other decays into hadrons and one neutrino (τ_had). By removing the muon information from the signals used for reconstruction and identification of the τ_had candidate in the boosted pair, the efficiency is raised to the level expected for an isolated τhad. The new procedure is validated by selecting a sample of highly boosted Z → τ_μτ_had candidates from the data sample of 140 fb-¹ of proton–proton collisions at 13 TeV recorded with the ATLAS detector. Good agreement is found between data and simulation predictions in both the Z → τ_μτ_had signal region and in a background validation region. The results presented in this paper demonstrate the effectiveness of the τ_had reconstruction with muon removal in enhancing the signal sensitivity of the boosted τ_μτ_had channel at the ATLAS detector.

Measurements of jet substructure are key to probing the energy frontier at colliders, and many of them use track-based observables which take advantage of the angular precision of tracking detectors. Theoretical calculations of track-based observables require ‘track functions’, which characterize the transverse momentum fraction r_q carried by charged hadrons from a fragmenting quark or gluon. This letter presents a direct measurement of r_q distributions in dijet events from the 140 fb⁻¹ of proton–proton collisions at √s=13 TeV recorded with the ATLAS detector. The data are corrected for detector effects using machine-learning methods. The scale evolution of the moments of the r_q distribution is sensitive to non-linear renormalization group evolution equations of QCD, and is compared with analytic predictions. When incorporated into future theoretical calculations, these results will enable a precision program of theory-data comparison for track-based jet substructure observables.

This paper reports the measurement of Higgs boson production in association with a pair in the decay channel. The analysis uses 140^-1 fb of 13 TeV proton–proton collision data collected with the ATLAS detector at the Large Hadron Collider. The final states with one or two electrons or muons are employed. An excess of events over the expected background is found with an observed (expected) significance of 4.6 (5.4) standard deviations. The cross-section is for a Higgs boson mass of 125.09 GeV, consistent with the prediction of the Standard Model of fb. The cross-section is also measured differentially in bins of the Higgs boson transverse momentum within the simplified template cross-section framework.

The Lund jet plane (LJP) is measured for the first time in events, using 140^-1  of √s = 13 TeV pp collision data collected with the ATLAS detector at the LHC. The LJP is a two-dimensional observable of the sub-structure of hadronic jets that acts as a proxy for the kinematics of parton showers and hadron formation. The observable is constructed from charged particles and is measured for R = 1.0 anti-k_t jets with transverse momentum above 350 GeV containing the full decay products of either a top quark or a daughter W boson. The other top quark in the event is identified from its decay into a b-quark, an electron or a muon and a neutrino. The measurement is corrected for detector effects and compared with a range of Monte Carlo predictions sensitive to different aspects of the hadronic decays of the heavy particles. In the W-boson-initiated jets, all the predictions are incompatible with the measurement. In the top quark initiated jets, disagreement with all predictions is observed in smaller subregions of the plane, and with a subset of the predictions across the fiducial plane. The measurement could be used to improve the tuning of Monte Carlo generators, for better modelling of hadronic decays of heavy quarks and bosons, or to improve the performance of jet taggers.

The mass of the top quark is measured using top-quark-top-antiquark pair events with high transverse momentum top quarks. The dataset, collected with the ATLAS detector in proton–proton collisions at √s=13 TeV delivered by the Large Hadron Collider, corresponds to an integrated luminosity of 140 fb⁻¹. The analysis targets events in the lepton-plus-jets decay channel, with an electron or muon from a semi-leptonically decaying top quark and a hadronically decaying top quark that is sufficiently energetic to be reconstructed as a single large-radius jet. The mean of the invariant mass of the reconstructed large-radius jet provides the sensitivity to the top quark mass and is simultaneously fitted with two additional observables to reduce the impact of the systematic uncertainties. The top quark mass is measured to be m_t=172.95±0.53 GeV, which is the most precise ATLAS measurement from a single channel.

This paper presents measurements of top-antitop quark pair () production in association with additional b-jets. The analysis utilises 140 fb⁻¹ of proton–proton collision data collected with the ATLAS detector at the Large Hadron Collider at a centre-of-mass energy of 13 TeV. Fiducial cross-sections are extracted in a final state featuring one electron and one muon, with at least three or four b-jets. Results are presented at the particle level for both integrated cross-sections and normalised differential cross-sections, as functions of global event properties, jet kinematics, and b-jet pair properties. Observable quantities characterising b-jets originating from the top quark decay and additional b-jets are also measured at the particle level, after correcting for detector effects. The measured integrated fiducial cross-sections are consistent with  predictions from various next-to-leading-order matrix element calculations matched to a parton shower within the uncertainties of the predictions. State-of-the-art theoretical predictions are compared with the differential measurements; none of them simultaneously describes all observables. Differences between any two predictions are smaller than the measurement uncertainties for most observables.