betaFeedback

WorksPhysics and Astronomy

A fully differential charged-particle phase-space measurement of ZZ+jets production in pppp collisions at s=13\sqrt{s}=13 TeV with the ATLAS detector

Abstract

The production of a ZZ~boson decaying to muons in association with hadronic jets is a benchmark process in proton-proton collisions at the Large Hadron Collider. It provides a clean experimental signature with high selection efficiency and purity while probing a wide range of Standard Model dynamics. Measurements of this process are typically reported as binned differential cross sections in a limited set of predefined observables, which reduces the information retained and hinders reinterpretation. In this paper, the OmniFold method is used to produce an unbinned and highly differential measurement of the ZZ+jets cross section using 140 fb−1\text{fb}^{-1} of s=13\sqrt{s} = 13 TeV proton-proton collision data collected by the ATLAS detector. This constitutes the first published unbinned, fully differential measurement of a collider cross section, describing the kinematics of individual charged particles within a well-defined region of phase space. This unbinned event representation is used to construct measurements in four example applications that illustrate the broad range of observables constrained with a single measurement, including several that would be difficult or impossible to access using traditional unfolding methods. The measurement is published as particle-level Monte Carlo events together with unfolded event weights, allowing arbitrary observables to be constructed after publication and enabling a broad range of future applications and reinterpretations.

© 2026 CERN for the benefit of the ATLAS Collaboration.

Reproduction of this article or parts of it is allowed as specified in the CC-BY-4.0 license.

Introduction

Precision differential cross-section measurements of Standard Model (SM) processes at the Large Hadron Collider (LHC) play a central role in contemporary particle physics. Such measurements provide stringent tests of theoretical predictions, constrain free parameters of the SM, and can reveal deviations associated with physics beyond the Standard Model (BSM). Measurements of the production of a ZZ boson decaying to two muons are widely used for these purposes [1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17]. In events where the ZZ boson is produced with large transverse momentum (pTp_{\mathrm{T}}), the dimuon system recoils against a collimated spray of hadrons, referred to as a jet. By using the two muons for both triggering and event selection, the ZZ boson can be used to select a clean and largely unbiased sample of high-pTp_{\mathrm{T}} jets. Measurements of these jets provide powerful probes of the dynamics of the strong interaction at high energies [18, 19]. They also provide important inputs to precision quantum chromodynamics (QCD) calculations and to the tuning of Monte Carlo (MC) event generators [20].

An essential component of collider measurements is correcting for the effects of finite detector resolution and acceptance on the measured distributions. In most current differential cross-section measurements, this procedure, known as unfolding, is performed by binning the data into histograms using a list of predefined observables of interest and correcting the resulting histogram counts using the simulated detector response [21, 22, 23, 24, 25, 26]. The resulting measurement is therefore reported as a collection of differential cross sections in fixed histogram bins. While highly successful, this approach performs an irreversible compression of the data before detector effects are corrected for statistically. Binned unfolding methods scale poorly with dimensionality, making simultaneous measurements of many observables computationally and statistically challenging. Moreover, the choice of observables and binning scheme are fixed, so most analyses focus on a small set of observables chosen before the measurement is performed. Information not captured by those choices is discarded, limiting reinterpretation and preventing the construction of new observables after publication. There are also classes of observables that are difficult or impossible to access within the traditional framework, including statistical moments of differential cross sections and observables for which no practical response matrix can be constructed.

To overcome these limitations, a new class of unfolding methods based on artificial intelligence (AI) techniques was developed [27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49, 50]. One such method, OMNIFOLD [27, 51], is used in several measurements by experimental collaborations [4, 17, 52, 53, 54, 55, 56, 57, 58, 59]. OMNIFOLD produces an unbinned representation of the unfolded cross section by assigning weights to particle-level simulated events, allowing observables to be constructed after unfolding without requiring a separate unfolding procedure for each observable. In most existing applications, however, the unfolded event sample is subsequently projected onto a limited set of observables and binned for visualization and interpretation; see Ref. [60] for a review and discussion of unbinned measurements. A recent ATLAS measurement of Z+Z+jets events, for example, used OMNIFOLD to perform a simultaneous differential cross-section measurement of 24 observables [4]. This will hereafter be referred to as the MULTIFOLD analysis.

A particularly powerful application of AI-based unfolding methods is the ability to operate directly on the kinematics of final-state particles, rather than on a predetermined set of summary observables. Because the number of particles produced in each collision varies event-by-event, averaging about 50 charged particles and sometimes exceeding 250 in this measurement, the resulting phase space is both high-dimensional and variable-dimensional. Unlike traditional response-matrix techniques, AI-based unfolding methods can learn detector-response corrections directly in high-dimensional variable-length event representations. Rather than publishing a finite collection of differential cross sections, the primary result is an unfolded representation of the event-level phase space from which differential cross sections and other can subsequently be derived by summing the appropriate event weights. In the case of Omnifold, the published measurement consists of particle-level simulated events together with a nominal unfolded event weight and associated alternative weight variations that encode the measurement uncertainties. The nominal event weights are normalized to the measured fiducial cross section and therefore have units of femtobarns (fb). Consequently, the measured cross section corresponding to any particle-level selection can be obtained simply by summing the event weights of the selected events.

The execution of this measurement requires extending Omnifold to operate on variable-dimensional event representations using transformer neural networks, together with a large-scale pretraining procedure on simulation to achieve sufficient precision in the vastly enlarged phase space. To enable direct validation, the simulated event samples, event selection, and target phase space are identical to those used in the Multifold analysis, allowing comparisons with the 24 observables measured previously while simultaneously extending the measurement to substantially higher dimensionality. By measuring the kinematics of the individual charged particles produced in each collision, a large number of observables are implicitly measured by a single cross-section measurement [61]. The measurement therefore shifts the experimental focus from publishing observables to publishing an event-level representation from which observables can be constructed. This includes observables whose measurement would be prohibitively complex or impossible to access within conventional binned unfolding frameworks. For example, traditional measurements of jet substructure observables require careful calibration of jets built with a particular clustering algorithm [62]. With fully differential charged-particle phase-space measurements, any jet clustering algorithm can be applied to the particle-level event representation, and valid measurements of jet observables can be constructed from the resulting jets without the need for additional calibration. This provides access to measurements of a much broader range of jet observables than are currently practically accessible to particle physics experiments. Because the measurement remains unbinned, observables can be rebinned to suit downstream analyses and statistical moments can be computed without introducing binning effects. Several examples of these applications are presented in this paper. Specifically, § 5.2, § 5.3, and § 5.4 present measurements of jet-shape observables defined with multiple jet clustering algorithms, energy correlators, and the intrinsic dimensionality of jet data, respectively, all derived from the same unfolded charged-particle event sample. A limitation of the present measurement is that it constrains only the charged-particle component of the final state, since robust systematic uncertainties for neutral-particle measurements are not currently available [63, 64].

Given the breadth of possible applications, this paper focuses on four representative use cases introduced in § 5. To encourage further study and reinterpretation, the published measurement is released publicly on the CERN open data portal [65]. This release enables the measurement of any observable that is a function of the measured phase space. All physics results presented in this paper are derived from this single underlying measurement. Additionally, a user guide with detailed examples and usage recommendations is provided to accompany the data [66].

Experimental setup

This measurement is performed using the full Run 2 proton–proton collision dataset collected by the ATLAS detector [67] at s=13\sqrt{s} = 13 TeV, corresponding to an integrated luminosity of 140140 fb−1^{-1} [68]. The ATLAS experiment is a multipurpose particle detector with a forward–backward symmetric cylindrical geometry and a near 4π4\pi coverage in solid angle.1 It consists of an inner tracking detector surrounded by a thin superconducting solenoid providing a 2 T axial magnetic field, electromagnetic and hadronic calorimeters, and a muon spectrometer.

The ATLAS inner tracking detector covers the pseudorapidity range ∣η∣<2.5|\eta| < 2.5. It consists of silicon pixel, silicon microstrip, and transition radiation tracking detectors. Lead/liquid-argon (LAr) sampling calorimeters provide electromagnetic (EM) energy measurements with high granularity within the region ∣η∣<3.2|\eta| < 3.2. A steel/scintillator-tile hadronic calorimeter covers the central pseudorapidity range (∣η∣<1.7|\eta| < 1.7). The endcap and forward regions are instrumented with LAr calorimeters for EM and hadronic energy measurements up to ∣η∣=4.9|\eta| = 4.9. The muon spectrometer surrounds the calorimeters and is based on three large superconducting air-core toroidal magnets with eight coils each. The muon spectrometer includes a system of precision tracking chambers up to ∣η∣=2.7|\eta| = 2.7 and fast detectors for triggering up to ∣η∣=2.4|\eta| = 2.4. The luminosity is measured mainly by the LUCID–2 [69] detector which is located close to the beampipe. A two-level trigger system was used to select events [70]. The first-level trigger is implemented in hardware and uses a subset of the detector information to accept events at a rate close to 100 kHz. This is followed by a software-based trigger that reduces the accepted rate of complete events to 2 kHz on average depending on the data-taking conditions. A software suite [71] is used in data simulation, in the reconstruction and analysis of real and simulated data, in detector operations, and in the trigger and data acquisition systems of the experiment.

Particle-level phase-space definition

The signal is defined as events containing a Z→μμZ \to\mu\mu decay. The particle-level phase space, which defines the cross section being measured, consists of final-state charged particles with a mean lifetime satisfying cτ0>10c\tau_0 > 10 mm. The fiducial region is defined by the following requirements. Events are required to contain at least two muons that do not originate from hadron decays and must each satisfy transverse momentum pT>25p_{\mathrm{T}} > 25 GeV and pseudorapidity ∣η∣<2.4|\eta| < 2.4. The two highest-pTp_{\mathrm{T}} muons are selected and required to have opposite charge, dimuon invariant mass mμμ∈(81,101)m_{\mu\mu} \in(81, 101) GeV, and dimuon transverse momentum pTμμ>200p_{\mathrm{T}}^{\mu\mu} > 200 GeV. The requirement pTμμ>200p_{\mathrm{T}}^{\mu\mu} > 200 GeV selects the boosted regime, where the recoiling hadronic system contains energetic jets without imposing any explicit selection on the jets themselves. This provides an unbiased sample for studies of charged-particle jet substructure while maintaining a high signal purity and sufficient statistical precision for the measurement. All such charged particles with pT>0.5p_{\mathrm{T}} > 0.5 GeV and ∣η∣<2.5|\eta| < 2.5 are included in the particle-level event representation.

Each particle-level event is represented by the set of charged particles satisfying these requirements, where each particle is described by its kinematic properties (pT,η,ϕ,m)(p_{\mathrm{T}, \eta, \phi, m}). At particle level, the mass assigned to each charged particle corresponds to its particle species. This information is not constrained by the unfolding procedure and is simply taken from the MC simulated samples used to construct the measurement (see § 3.6). Because the number of particles varies from event to event, the resulting event representation is variable in length.

Data and simulated event samples

Data events are collected using single-muon triggers [72]. Muons are reconstructed by matching charged-particle tracks in the inner tracking detector and muon spectrometer, accounting for energy loss in the calorimeters [73]. They are required to fulfill Medium identification criteria and PflowLoose isolation [73], and must satisfy ∣d0∣/σd0<3\lvert d_0\rvert/\sigma_{d_0} < 3 and ∣z0sin⁡(θ)∣<0.5\lvert z_0 \sin(\theta)\rvert< 0.5 mm. The transverse impact parameter d0d_0 is the signed transverse distance from the beamline to the point of closest approach of the muon track, while the longitudinal impact parameter z0z_0 is the difference between the longitudinal coordinate of this point and that of the reconstructed primary vertex. These requirements ensure that the selected muons originate from the interaction vertex, i.e. the primary vertex with the highest sum of associated track pT2p_{\mathrm{T}}^2 [74]. Charged-particle tracks used for jet building are required to satisfy the Loose track-quality criteria, the Tight track-to-vertex association criterion to suppress tracks originating from pileup interactions [75], and must not be associated with the selected muons. Identical kinematic requirements are applied to the reconstructed quantities as those used to define the particle-level fiducial volume.

MC simulated event samples are used to provide theoretical predictions at both the reconstructed and particle level, and to perform the unfolding. Most Z+Z+jets events that satisfy the selection originate from the Drell–Yan process. As in the Multifold analysis, the nominal Monte Carlo prediction is provided by [email protected] [76, 77, 78] interfaced to Pythia 8.240 [79] (denoted MG5+Py8 in the following), and Sherpa 2.2.11 [78, 80] provides an alternative. Both the nominal and alternative Monte Carlo predictions use the NNPDF3.0nnlo parton distribution function (PDF) set [81]. Electroweak Z+Z+jets events are simulated with Herwig 7.2.0 [82, 83], interfaced with VBFNLO v3.0.0 [84] using the MMHT2014LO PDF set [85]. Diboson production is generated at next-to-leading-order accuracy with Sherpa 2.2.1 or 2.2.2 [80], employing the NNPDF3.0nnlo PDF set [81]. Background contributions from top processes (ttˉt\bar{t}, tVtV, single top) are modeled by Powheg v2 [86] interfaced to Pythia 8.230, and an alternative description of the ttˉt\bar{t} background is provided by Sherpa 2.2.1 [80]. Both the nominal and alternative descriptions of the ttˉt\bar{t} background utilize the NNPDF3.0nnlo PDF set [81]. The detector response is simulated using a dedicated Geant4-based model [87] of the ATLAS detector [88]. Simulated inelastic minimum-bias events are overlaid to model additional pppp collisions in the same and neighboring bunch crossings (“pileup”) [88]. Simulated events are reconstructed using the same procedure as for data.

The analysis measures inclusive Z+Z+jets production and makes no attempt to separate the Drell–Yan, diboson, and electroweak processes. Although the measurement does not explicitly separate these processes, the released particle-level simulated samples contain a per-event label that makes it possible to obtain individual predictions of each of the components. See Appendix § 8.1 for a discussion of proper procedures for these use cases. The resulting inclusive Z+Z+jets signal sample consists of approximately 95.3% Drell–Yan, 2.9% diboson (primarily ZV→μμjjZV \to\mu\mu jj), and 1.6% electroweak ZjjZjj production. These samples are used as the starting point for the top background subtraction and unfolding procedures described below.

Analysis methodology

This section describes the methodology used to construct the unbinned measurement of the charged-particle final-state phase space. The measurement is obtained through a sequence of steps including background subtraction, iterative unfolding with Omnifold, and the assignment of unfolded weights to particle-level MC events. These steps are described below, along with additional components of the methodology including the event representation, neural-network architecture, pretraining procedure, and generation of the pseudodata sample used in the simulation-based validation.

Background subtraction

Top-quark processes are expected to account for approximately 0.25% of the selected event yield and constitute the only non-negligible background in the fiducial region. Prior to unfolding, this small residual background is removed from the data using an unbinned subtraction procedure. In the Multifold measurement of the same process, explicit subtraction of the top-quark background was deemed unnecessary given its sufficiently small contribution. Instead, its estimated size was assigned as a systematic uncertainty. In the present analysis, the expected background is directly subtracted as part of the measurement procedure rather than being treated as a systematic uncertainty, thereby reducing the overall measurement uncertainty.

The subtraction is performed using a neural-network-based reweighting method [89]. In principle, the reweighting could be derived in the full event phase space used for the unfolding. However, because the top-quark contribution is extremely small, it was found that a sufficiently accurate reweighting could not be achieved using the full high-dimensional event representation. Instead, the subtraction is performed in a reduced space consisting of five observables, including the output of a classifier trained on the full charged-particle phase space to distinguish ttˉ\mathrm{t\bar{t}} events from Z+jets events. Incorporating this highly sensitive discriminator improves the ttˉ\mathrm{t\bar{t}} subtraction throughout the charged-particle phase space, even in observables that were not explicitly used in the subtraction procedure. The subtraction procedure assigns an event weight to each data event, yielding a statistically background-subtracted data sample that is subsequently used as input to the Omnifold procedure. See Appendix § 10.2 for additional details and results of the subtraction procedure.

The Omnifold method

Omnifold is an iterative unfolding algorithm that uses machine-learning classifiers to estimate likelihood ratios between the probability distributions describing data and MC simulation. The final result is a set of weights assigned to particle-level simulated events, which together form an unbinned representation of the unfolded cross section.

Schematics illustrating the Omnifold procedure and the variable-length event representation are shown in Figure 1. One iteration of the procedure estimates the detector-level and particle-level reweighting functions ω(x⃗r)\omega(\vec{x}_{r}) and ν(x⃗p)\nu(\vec{x}_{p}), which depend on the detector-level and particle-level event representation, x⃗r\vec{x}_{r} and x⃗p\vec{x}_{p}, respectively. The detector-level function ω(x⃗r)\omega(\vec{x}_{r}) reweights simulated events to reproduce the observed data distribution. These weights are then propagated through the detector simulation to the corresponding particle-level events, which are used to train the particle-level function ν(x⃗p)\nu(\vec{x}_{p}). The resulting particle-level weights define the particle-level MC prior for the next iteration. The procedure is repeated over multiple iterations, yielding the final Omnifold weights νo(x⃗p)\nu^{\mathrm{o}}(\vec{x}_{p}). As in the Multifold analysis, the unfolded event weights are subsequently normalized such that their sum reproduces the measured fiducial cross section obtained from the data event count, luminosity, efficiency, and fiducial acceptance. Consequently, the published event weights have units of femtobarns (fb), and the measured cross section corresponding to any particle-level selection is obtained by summing the weights of the selected events. The overall unfolding procedure is an example of an expectation-maximization algorithm and is formally the zero bin-width limit of the commonly used Iterative Bayesian Unfolding (IBU) method [24].

Schematic illustrating the Omnifold procedure and the variable-dimensional nature of the fully differential charged-particle phase-space measurement

Figure 1. Schematic illustrating (a) the Omnifold procedure and (b) the variable-dimensional nature of the fully differential charged-particle phase-space measurement. Omnifold starts by deriving a reweighting function ω(x⃗r)\omega(\vec{x}_r) of the detector-level MC events x⃗r\vec{x}_r to match the data distribution (step 1). This reweighting function is then propagated from detector level to particle level, and a second reweighting function ν(x⃗p)\nu(\vec{x}_p) of the particle-level MC events x⃗p\vec{x}_p is derived to match the raw particle-level MC distribution to the ω(x⃗r)\omega(\vec{x}_r)-adjusted particle-level MC distribution (step 2). The result of the second step is then propagated from particle level back to detector level and the process repeats for a total of seven iterations. The final result of the procedure is the particle-level reweighting function νof(x⃗p)\nu^{\mathrm{of}}(\vec{x}_p) after seven iterations. This reweighting function is used to produce an event weight for each event in the nominal particle-level MC sample. These event weights are well defined despite the variable dimensionality because the neural networks are set-processing transformer architectures, allowing observables such as the number of charged hadrons, nchn_{\mathrm{ch}}, to be measured even though they are not explicitly provided to the networks. The first two events in the published dataset contain 51 and 37 charged particles, corresponding to variable-length event representations with 153 and 111 dimensions, respectively.

The unfolding is performed using seven iterations. This choice was determined using the pseudodata validation procedure described in § 3.5. The agreement between the unfolded pseudodata measurement and the known particle-level pseudodata target is quantified using a distance metric defined as the sum of one-dimensional Wasserstein distances [90] computed over many event-level observables. No significant improvement was observed beyond seven iterations. The unfolding is derived in a slightly enlarged fiducial region (pμμ>190p^{\mu\mu} > 190 GeV) to reduce sensitivity to migrations across the fiducial boundary. Once derived, the unfolded weights associated with each particle-level event are fixed and can be reused to construct arbitrary observables without retraining or rerunning the unfolding procedure.

The published measurement is constructed from a statistically independent test subset of the nominal MC prediction. The effective event yield of this test set is equal to approximately 2.5 times that of the data.

This test subset is used to define the published measurement and does not participate in the neural-network training or in the determination of the final reweighting function νo(x⃗p)\nu^{o}(\vec{x}_{p}).

Variable-length event representation and neural-network architecture

Unlike the MULTIFOLD analysis, where the unfolded weights define cross sections only in a fixed set of 24 observables, the present measurement assigns weights directly to particle-level event configurations and therefore provides a measurement of the full charged-particle final-state phase space. This is achieved by replacing the multi-layer perceptron architectures used by the MULTIFOLD analysis with Particle Transformer (ParT) networks [91], which can operate directly on variable-length particle clouds. Each event is represented as a variable-length particle cloud consisting of the four-vectors (pT,η,ϕ,m)(p_{\mathrm{T}}, \eta, \phi, m) of all charged particles. At both detector and particle level, the mass of the charged particles is set to the pion mass, except for the muons produced in the ZZ boson decay, which are assigned the muon mass. This representation differs from that used in defining the fiducial volume in § 2.1 and used in subsequent analyses of the measurement in § 5, where all charged particles are assigned their particle-level masses. This design choice was made to prevent the reweighting produced by OMNIFOLD from altering the fractions of particle-level hadron identities, which are not constrained at detector level. Instead, the measurement assumes the particle-level hadron fractions predicted by the MG5+Py8 generator.

In addition to the particle four-vectors, each particle is assigned the index of the jet to which it belongs. These jet indices are calculated by clustering all particles with the anti-kTk_{\mathrm{T}} algorithm with a radius parameter of R=0.4R = 0.4. Particles associated with jets beyond the fifth-highest-pTp_{\mathrm{T}} jet are grouped into a single additional category. These six classes of particles are one-hot encoded and provided as additional features to the network. These auxiliary features are fully determined by the three-vectors (pT,η,ϕ)(p_{\mathrm{T}}, \eta, \phi) of all particles, but were found to improve the accuracy of the likelihood ratios predicted by the neural networks. Events passing the fiducial selection contain, on average, about 50 charged particles, implying that the unfolding is performed simultaneously in roughly 3×50=1503 \times50 = 150 kinematic coordinates. Charged-particle multiplicities can exceed 250, corresponding to more than 750 simultaneous kinematic coordinates.

Pretraining and optimization

About 247 000 data events pass the detector-level selection; as a result, the first-stage classification problems within OMNIFOLD are strongly statistics limited. To address this limitation, a large-scale pretraining procedure is employed in which neural networks are trained from scratch on a large sample of simulated collision events, then fine-tuned on the density ratio estimation tasks within the OMNIFOLD procedure. Specifically, networks are pretrained to distinguish Drell–Yan events generated with MG5+Py8 from those generated with SHERPA. Dedicated subsets of the simulated event samples from both generators are used for pretraining and not used in the following OMNIFOLD procedure. In total, approximately 14 million events are used in the pretraining. Ten pretrained networks are trained independently from scratch for both the detector-level and particle-level phase spaces.2 These pretrained networks are subsequently used as warm-start checkpoints for the neural-network training within OMNIFOLD, with one of the ten checkpoints corresponding to the appropriate phase space selected at random for each training. The pretraining procedure was found to substantially reduce both the method bias observed in pseudodata studies and the variance between independent unfolding runs. It was an essential addition to the Omnifold procedure that enabled an accurate unfolding in an O(100)O(100)-dimensional phase space. This finding motivates additional development of large-scale pretrained foundation models for particle physics data [92, 93, 94, 95, 96, 97, 98, 99, 100, 101], which could be valuable tools for additional AI-based unfolding and other simulation-based inference applications [102, 103] in the future.

The ParT networks are implemented in PyTorch [104] and trained on the Perlmutter supercomputer at NERSC using between 16 and 64 NVIDIA A100 GPUs. A ParT model with six attention blocks, one class attention block, and a dropout probability of 0.02 applied to all attention weights provided the best performance in the validation studies. These networks contain approximately 1.6 million trainable parameters and were trained using the Lion optimizer [105] with a weight decay parameter of 0.01 and a cosine-annealing learning-rate schedule. The network architecture and training hyperparameters were determined through an extensive random hyperparameter search. Details of the hyperparameter optimization procedure and the final hyperparameter choices are provided in Appendix § 10.3.

Construction of pseudodata event samples

The primary validation of the measurement is performed using pseudodata generated from statistically independent simulated event samples. The dominant Drell–Yan contribution, comprising approximately 95% of the selected events, is modeled using Sherpa rather than the nominal MG5+Py8 prediction, thereby providing an independent description of the hard scattering, parton shower, and hadronization. The smaller electroweak ZjjZjj and diboson contributions are generated using the same event generators as in the nominal prediction, but with statistically independent simulated event samples.

To avoid potential circularity in the validation, the pseudodata are not constructed using neural-network-based reweighting methods. Instead, the detector-level simulated events are reweighted to resemble the observed data using an independent sequence of Gaussian-kernel reweighting functions. These smooth reweighting functions are derived sequentially from one-dimensional projections of event-level observables, including the 24 observables measured by the Multifold analysis, together with additional observables that constrain soft hadronic activity, such as charged-hadron multiplicities in different regions of transverse momentum and pseudorapidity, and the scalar sum of the transverse momenta of all charged particles in each event. The reweighting functions are derived and applied iteratively until good agreement with the data is achieved across this broad set of observables.

The reweighted detector-level events are statistically downsampled to match the observed event count in data and used as pseudodata. The corresponding particle-level events, which retain their full statistical precision and have an effective event yield approximately 12 times that of the data, provide the particle-level target against which the unfolded measurement is compared. These pseudodata samples form the basis of the validation studies described in § 4.3.

Considerations for using the measurement

Several considerations should be kept in mind when using the unbinned measurement. First, a measurement of a given observable should be validated using the pseudodata measurement before drawing any conclusion based on the measurement with real data. This is particularly important because it is possible to define observables using the particle-level phase space that are not constrained by the measurement, for example those sensitive to the particle-level hadron fractions predicted by the MG5+Py8 generator. Second, the unfolding process assumes that a sufficient number of events fall in each bin of the particle-level histogram, so a binning should be chosen that ensures at least 5000 effective events are in each bin. Finally, all events are guaranteed to contain two muons given the event selection, but for a small fraction of events, most charged-hadron tracks fall outside the ∣η∣<2.5\lvert\eta\rvert< 2.5 acceptance, resulting in minimal hadronic activity in the fiducial volume. When studying jet observables it is advisable to remove these events with a small minimum charge-hadron-jet pTp_{\mathrm{T}} requirement. These considerations are discussed in more detail in the documentation accompanying the public codebase and in Appendix § 8.1, which provide practical recommendations for using the measurement.

Uncertainties and validation

The unbinned measurement is defined by a nominal set of particle-level events and their associated unfolded weights. Uncertainties are propagated through the full analysis chain and represented by alternative realizations of the measurement, corresponding to alternative sets of unfolded particle-level event weights. These alternative weight sets can be projected onto arbitrary observables and combined to construct covariance matrices in the same manner as for conventional differential cross-section measurements.

In total, the measurement is accompanied by 162 weight variations that quantify its uncertainties. Twenty-seven weight variations encode independent systematic uncertainty sources. The remaining 135 weight variations encode stochastic uncertainties arising from the finite statistical precision of the data and simulated event samples. These are evaluated using bootstrap replicas [106].

Experimental and modeling uncertainties

Experimental sources of uncertainty include systematic bias due to the muon efficiency and calibration [73], track reconstruction [75], pileup modeling, and the luminosity measurement [68]. Theoretical uncertainties in the Drell–Yan component of the MC are evaluated for variations of PDF and αs\alpha_{\mathrm{s}} choices [78], QCD scales [78], and the generator tune [107]. To approximate the expected impact of theoretical uncertainties on the small contributions from diboson and electroweak ZjjZjj processes, flat yield variations of 30% and 20% are applied to the predicted yield of each source, respectively, corresponding to conservative estimates of the level of agreement to data as observed in Refs. [108] and [109]. The uncertainty associated with the top-background subtraction is evaluated by repeating the subtraction procedure using the alternative \textsc{Sherpa} ttˉt\bar{t} prediction in place of the nominal model. The resulting alternative subtraction weights are propagated through the full \textsc{Omnifold} procedure, and the difference relative to the nominal measurement is assigned as an uncertainty.

The relative importance of the individual uncertainty sources depends on the observable under consideration. For observables primarily sensitive to the muon kinematics, the dominant experimental uncertainty arises from the muon isolation and reconstruction efficiencies. In contrast, for observables that depend strongly on charged-particle kinematics, such as the jet kinematic observables, the charged-hadron multiplicity (nchn_{\mathrm{ch}}), and the scalar sum of charged-hadron transverse momenta (HTH_{\mathrm{T}}), the dominant experimental uncertainty is typically associated with the track reconstruction efficiency. Across most measured observables, the experimental uncertainties exceed the theoretical modeling uncertainties, with the latter generally contributing at a level comparable to or smaller than the luminosity uncertainty [68]. The remaining uncertainty sources are associated with the unfolding procedure itself and are discussed in the following subsection.

Unfolding uncertainties

In addition to the experimental and theoretical uncertainties described above, several uncertainty sources associated with the unfolding procedure itself are evaluated: training stochasticity, data-driven finite-iteration uncertainty, hadron-composition dependence, and hidden-variable dependence.

Neural-network optimization is inherently stochastic and yields slightly different unfolded weights for different training realizations. This uncertainty is estimated by repeating the Omnifold procedure with identical inputs but different random seeds governing the training and validation splits, as well as minibatch construction. In contrast to [4], ten repetitions are used instead of one hundred because the pretraining procedure described in § 3.3 substantially reduces the variance between independent training realizations while maintaining the accuracy of the unfolded results. The resulting spread in the unfolded weights is assigned as the stochastic training uncertainty.

Like IBU, Omnifold returns a biased estimator of the true particle-level distribution. The residual bias arising from the finite number of iterations is evaluated using a data-driven procedure. The MG5+Py8 sample used for the nominal measurement is reweighted at particle level such that the corresponding detector-level distributions reproduce the observed data. To ensure independence from the neural-network-based reweighting used by Omnifold, the same Gaussian-kernel-based reweighting procedure described in § 3.5 is used. The resulting reweighted sample is then treated as pseudodata and unfolded using the nominal procedure. The difference between the unfolded result and the known particle-level target distribution is assigned as the data-driven finite-iteration uncertainty.

The unfolded result can also be biased by latent features in the simulation of the detector response that are not included as input variables. The most significant of these hidden variables are the particle-level hadron fractions predicted by the MC sample used to derive the unfolding. Measurements of the kinematic properties of charged hadrons have different biases and resolutions depending on the hadron type, so changing the charged hadron composition will change the detector response. To assess potential biases in the unfolded results arising from differences in hadron composition, the hadron content of the nominal MG5+Py8 sample is reweighted at particle level to match that of the Sherpa sample. Four hadron categories are defined for this procedure: pions, kaons, protons, and an ‘other’ category that includes any charged hadron that is not identified as a pion, kaon, or proton. The majority of hadrons in this category are strange baryons, predominantly Ω\Omega, Ξ\Xi, and Σ\Sigma baryons. The reweighting is performed to match several key distributions within each category to the Sherpa sample. These include the distribution of the hadron multiplicity fraction, the distribution of the fraction of jet momentum carried by each hadron category in the leading and subleading jets, and the distribution of the fraction of the total event transverse momentum carried by each hadron category when excluding tracks inside the leading and subleading jets. The unfolding is then repeated with the reweighted MG5+Py8 sample, and the difference from the nominal result is taken as the hadron composition uncertainty.

Even a highly differential unfolding may remain sensitive to latent detector-response variables that are not explicitly included in the event representation. To estimate the impact of such effects, the unfolding is repeated with the alternative Sherpa sample used as the MC sample, after it is reweighted using the same Gaussian-kernel-based methodology described in § 3.5 to match the MG5+Py8 sample at particle level. The difference between the resulting unfolded spectra and the nominal results derived with the MG5+Py8 sample is taken as the hidden variable uncertainty.3

The Omnifold measurement is considerably more differential than conventional binned unfolding methods and is therefore, in principle, sensitive to fewer hidden detector-response variables. Nevertheless, important latent variables remain. In particular, the particle-level hadron composition is not constrained by the detector-level event representation and therefore cannot be inferred directly from the observed data. Consequently, even a measurement of the fully differential charged-particle phase space retains a non-negligible hidden-variable uncertainty.

Compared with IBU, the hidden-variable uncertainty can be slightly larger for Omnifold, particularly in low-statistics regions, owing both to these remaining latent variables and to the finite accuracy of the neural-network likelihood-ratio estimates. Comparisons of the total uncertainty and hidden-variable uncertainty for the IBU, Multifold, and Omnifold measurements are provided in Appendices § 10.§ 2 and § 10.§ 3, respectively. For the 24 observables measured in the Multifold analysis, the unfolding uncertainty of the Omnifold measurement is generally comparable in magnitude to the leading experimental uncertainties across much of the fiducial phase space, although it can become significantly larger in low-statistics regions. It therefore constitutes one of the dominant contributions to the overall uncertainty budget.

Validation

The measurement is validated in multiple ways, with the resulting recommendations for downstream use of the published measurement summarized in Appendix § 8.1. The primary validation is based on an auxiliary measurement in which the pseudodata described in § 3.5 are treated as data and unfolded. The resulting pseudomeasurement is compared with the known particle-level pseudodata target using binned χ2\chi^2 goodness-of-fit tests across a broad range of observables, accounting for all sources of uncertainty and their bin-to-bin correlations through the full covariance matrix.

Figure 2(a) shows a representative example of this validation for nchn_{\mathrm{ch}}, one of the new observables constructed from the variable-length event representation and presented in § 5.1. Good agreement with the known particle-level pseudodata target is observed across the distribution. Appendix § 9.1 provides further details of the statistical methodology and summarizes the resulting χ2\chi^2 values, numbers of degrees of freedom, and p-values for the 24 observables measured in [4], together with two additional observables, nchn_{\mathrm{ch}} and HTH_{\mathrm{T}}, presented in § 5.1. The resulting p-values span a broad range and show no systematic indication of disagreement between the pseudomeasurement and the particle-level pseudodata target. The appendix also presents the full pseudodata results, including uncertainty breakdowns and correlation matrices, for these two additional observables. The performance of the method was further studied using pseudodata for a broader set of projections, including two-dimensional distributions and measurements in restricted kinematic regions, such as diboson- and vector-boson-fusion-enhanced regions.

Figure 2. Representative validation results. (a) Comparison of the unfolded pseudomeasurement (points) with the known particle-level target (triangles) for the charged-hadron multiplicity, nchn_{\mathrm{ch}}. The corresponding goodness-of-fit test yields χ2=2.41\chi^2 = 2.41 with a pp-value of 0.790. (b) Comparison of the differential cross section, dσ/dpTμμ\mathrm{d}\sigma/\mathrm{d}p_{\mathrm{T}}^{\mu\mu}, measured using IBU, Multifold from Ref. [4], and Omnifold (denoted “This result”).

An additional pseudodata validation targets regions in which the nominal MG5+Py8 sample predicts a much lower cross section than the Sherpa sample used to construct the pseudodata. A ParT-based classifier is trained to distinguish particle-level Drell–Yan events generated by MG5+Py8 and Sherpa using the event representation described in § 3.3. Since the classifier is optimized to identify differences between the two generator distributions, its output provides a sensitive observable for probing regions where the particle-level densities predicted by the two generators differ substantially.

The high-SHERPA-probability tail of the classifier output is binned following the statistical requirements of Appendix § 8.1. In this region, the MG5+Py8-based nominal event density is approximately an order of magnitude smaller than that of the Sherpa-based pseudodata target. Nevertheless, the unfolded pseudomeasurement is consistent with the known particle-level pseudodata target, with a χ2\chi^2 test yielding a pp-value of 0.863. Further details are provided in Appendix § 9.1. The unfolded event-weight distribution provides a complementary indication of the magnitude of the reweighting required in the measurement: only 0.01% of events have dimensionless weights νf\nu^f of greater than 50.

Additional validation is provided through comparisons with corresponding IBU-based measurements and the published Multifold measurement [4]. Figure 2(b) shows a representative comparison of the differential cross section measured using IBU, Multifold, and Omnifold for the transverse momentum of the dimuon system, pTμμp_{\mathrm{T}}^{\mu\mu}. The three measurements are consistent within their uncertainties across the full kinematic range. More extensive comparisons across the 24 observables measured in Ref. [4], together with the two additional observables presented in this paper, are provided in Appendix § 10. Together, these studies motivate the usage recommendations in Appendix § 8.1, which are followed for the results presented in the next section.

Results

The published measurement [65, 66] constitutes the primary result of this work. The trained networks produced by the Omnifold procedure are also released publicly, allowing appropriate unfolded event weights to be calculated for arbitrary samples generated with MG5+Py8 and Sherpa.

Because the published measurement is unbinned, high-dimensional, and supports a broad range of downstream applications, it is neither practical nor desirable to present an exhaustive set of projected observables in this paper. Instead, four representative applications are presented to illustrate how the

measurement can be used to construct observables spanning a wide range of physics topics, including event-level kinematics, jet substructure, energy-flow observables, and machine-learning-inspired characterizations of jet structure. The examples are arranged from conventional observables commonly measured in collider experiments to observables that are difficult or impossible to access using traditional unfolding techniques.

In all figures presented in this section, the data are shown with the full measurement uncertainties described in § 4. The measurements are compared with predictions for the sum of the Drell–Yan, electroweak ZjjZjj and diboson contributions. The dominant Drell–Yan component is modeled using either the nominal MG5+Py8 prediction or the alternative Sherpa prediction, while the electroweak ZjjZjj and diboson contributions are modeled using Herwig 7 and Sherpa, respectively. The uncertainties on the predictions include the experimental and modeling uncertainties described in § 4.1, with the dominant theoretical contributions arising from variations of the QCD scales and the generator tune.

Event-level observables

The most direct application of the full charged-particle phase-space measurement is the construction of conventional differential cross-section measurements by projecting the event-level representation onto observables of interest. Though such measurements can in principle be performed using traditional binned unfolding techniques, the event-level representation provides considerably greater flexibility for downstream analyses. Because the measurement is unbinned, additional event selections can be imposed after publication to define measurements in restricted regions of the fiducial phase space, provided sufficient statistical precision is maintained. Similarly, the binning of any observable can be modified on-the-fly to suit downstream applications such as Monte Carlo generator tuning. More generally, any observable that can be constructed from the unfolded muon and charged-particle kinematics is implicitly measured. This includes observables that were not considered during the analysis design and were not used as inputs to the unfolding procedure. The reinterpretation potential of the measurement therefore extends the long-term physics value of the ATLAS Run 2 dataset by enabling future measurements that do not require access to the detector simulation.

As representative examples, measurements of nchn_{\mathrm{ch}} and HTH_{\mathrm{T}} are shown in Figure 3. The muons from the ZZ boson decay are not included in either observable. Neither nchn_{\mathrm{ch}} nor HTH_{\mathrm{T}} enters the unfolding procedure directly; both are obtained entirely as projections of the published event-level representation, providing an independent demonstration of the predictive power of the published measurement. Validation studies using pseudodata are presented in Appendix § 9.1, including full uncertainty breakdowns and correlation matrices.

Figure 3. Measurements of (a) the number of charged hadrons nchn_{\mathrm{ch}} and (b) the scalar sum of the transverse momenta of all charged hadrons, HTH_{\mathrm{T}}. The measurement is shown as points, and particle-level predictions employing MG5+Py8 or Sherpa for the Drell–Yan contribution are shown as triangles and squares, respectively. Uncertainties on the measurements are shown as error bars and uncertainties on the predictions are shown as shaded bands.

The uncertainty bands associated with generator predictions reflect the modeling uncertainties described in § 4.1 and are dominated by the contribution arising from QCD scale variations in the Drell–Yan components. Both generator predictions exhibit significant discrepancies with the measured distributions. The largest discrepancies occur in the nchn_{\mathrm{ch}} distribution, where the simulation employing MG5+Py8 (Sherpa) for the Drell–Yan contribution predicts systematically too many (few) charged particles, with deviations reaching approximately 50% in some regions of phase space. The HTH_{\mathrm{T}} distribution is generally better described, although noticeable shape differences remain, particularly for the MG5+Py8-based prediction. These results demonstrate the ability of the measurement to constrain poorly modeled aspects of hadronic final states and provide valuable input for future tuning of Monte Carlo event generators.

Jet-shape observables

Jet-substructure observables provide sensitive probes of QCD radiation patterns and are widely used to study the internal structure of quark and gluon jets. In conventional measurements [110, 111, 112], jet observables are unfolded for a fixed jet definition, making systematic comparisons between clustering algorithms or radius parameters impractical. As the Omnifold measurement is differential in the kinematics of individual charged particles, jets can be built directly from the published event-level unfolded data using arbitrary clustering algorithms and parameters. This enables measurements of jet observables for many jet definitions without repeating the unfolding procedure.

As an example, Figure 4 shows a measurement of the average transverse-momentum density of charged hadrons as a function of distance from the jet axis, ⟨ρ(r)⟩\langle\rho(r) \rangle. Following Refs. [113, 114], the transverse-momentum density is defined by

Figure 4. Measurements of the average transverse-momentum density ⟨ρ(r)⟩\langle\rho(r)\rangle as a function of distance from the jet axis. (a): Comparison of Cambridge–Aachen R=0.4R = 0.4 jets between the unfolded measurement (points) and particle-level predictions with the Drell–Yan contribution modeled using MG5+Py8 or Sherpa (triangles and squares, respectively). Uncertainties on the measurement are shown as error bars and uncertainties on the generator predictions are shown as shaded bands. (b): Comparison of ⟨ρ(r)⟩\langle\rho(r)\rangle measured using jets clustered from the unfolded charged hadrons with three jet algorithms, all with a radius parameter of R=0.4R = 0.4: ktk_t (pentagons), anti-ktk_t (triangles), and Cambridge–Aachen (points). Uncertainties on the Cambridge–Aachen measurements are shown as error bars while uncertainties on ktk_t and anti-ktk_t measurements are shown as shaded bands.

⟨ρ(ri)⟩=⟨pT,iAi⟩jets(1)\left\langle\rho(r_i) \right\rangle= \left\langle\frac{p_{\mathrm{T},i}}{A_i} \right\rangle_{\mathrm{jets}} \tag*{(1)}

where pT,ip_{\mathrm{T},i} is the magnitude of the vector sum of the transverse momenta of all charged hadrons contained within a region in (y,ϕ)(y,\phi)-space with area AiA_i. The quantity rr describes the angular distance to the jet axis in Δy×Δϕ\Delta y \times\Delta\phi space, and each ii denotes an annulus region spanning a radial interval Δr\Delta r around rir_i. A given bin shown in Figure 4 covers the range r∈[ri−Δr,ri+Δr]r \in[r_i - \Delta r, r_i + \Delta r] with an annulus area of Ai=π(ri+Δr)2−π(ri−Δr)2A_i = \pi(r_i + \Delta r)^2 - \pi(r_i - \Delta r)^2. The angle brackets in Equation (1) denote the weighted average over the selected jet sample. Because the observable is defined as a weighted average over jets, it is insensitive to the overall normalization of the measured cross section. Only jets with pT∈[120,150]p_{\mathrm{T}} \in[120, 150] GeV and ∣y∣<1\lvert y \rvert< 1 are considered, as this interval provides a reasonably narrow selection of central jets located in the falling region of the leading jet pTp_{\mathrm{T}} distribution while retaining a large number of events.

Figure 4(a) illustrates a measurement of the average pTp_{\mathrm{T}} density of Cambridge–Aachen R=0.4R = 0.4 jets compared with theory predictions from the MG5+Py8 and Sherpa samples, shown as the triangular and square data points, respectively. Both generators tend to predict a higher transverse-momentum density close to the jet axis (r<0.04r < 0.04) and a lower density at intermediate radii (0.04<r<0.30.04 < r < 0.3), suggesting that the simulated jets are somewhat narrower than those observed in data. Figure 4(b) shows measurements of the same observable but using jets clustered by three different jet algorithms using the FastJet package [115]. While the Cambridge–Aachen and ktk_t jet algorithms produce similar results, the anti-ktk_t algorithm produces a large increase and subsequent decrease in pTp_T density relative to the other two algorithms near the jet boundary: r≈Rr \approx R. Although only results for R=0.4R = 0.4 are shown here, the same qualitative behavior is observed for other jet-radius parameters that were studied. This behavior can be understood as a consequence of the circular and greedy nature of the anti-ktk_t algorithm, which prioritizes high-pTp_T constituents during clustering. In the presence of a relatively uniform soft background, upward fluctuations near the jet boundary can influence the jet axis position, leading to an enhancement of the measured density just inside the nominal jet radius and a corresponding depletion near the boundary. This effect is less pronounced in jets built with the Cambridge–Aachen and ktk_t algorithms as the former relies on purely angular ordering, while the latter constructs jets beginning at low transverse momenta and so is not greedy. The corresponding enhancement just outside the nominal jet radius is a consequence of the same effect and is present for all three clustering algorithms in both the unfolded measurement and the simulations. Importantly, all three measurements shown in Figure 4(b) are derived from the same published event-level representation and require neither additional detector calibration nor repeated unfolding.

Historically, jet-shape observables measured in Z+jets events have attracted particular interest because the selected jets are predominantly quark initiated [116]. The ability to reconstruct jets directly from the unfolded particle-level event representation provides new opportunities for studying quark-jet properties at high pTp_T. More generally, additional selections based on observables constructed from the unfolded event representation can be applied a posteriori, enabling measurements in specialized phase-space regions without repeating the unfolding procedure. Examples include measurements in quark- and gluon-enriched jet samples, studies of alternative jet definitions, and investigations of novel jet-substructure observables that were not envisioned when the measurement was performed.

Energy correlator observables

Energy correlators are a class of observables that quantify energy-weighted correlations in particle collisions [117]. They can be calculated to high precision in perturbative QCD [118], are sensitive to the strong coupling constant αs\alpha_s [119, 120], and encode fundamental aspects of QCD radiation [121]. The simplest and most widely studied example is the two-point energy correlator (EEC), which measures the correlation between pairs of particles as a function of their opening angle. It is defined as

EEC⁡(z)=1σ∑i≠j∫dσET,iET,jQT2δ(z−1−cos⁡θij2).(2)\operatorname{EEC}(z) = \frac{1}{\sigma} \sum_{i \ne j} \int\mathrm{d}\sigma\frac{E_{\mathrm{T},i} E_{\mathrm{T},j}}{Q_{\mathrm{T}}^2} \delta\left(z - \frac{1-\cos\theta_{ij}}{2}\right). \tag*{(2)}

Here ET,iE_{\mathrm{T},i} and ET,jE_{\mathrm{T},j} are the transverse energies of the two objects, zz is the angular separation observable z=(1−cos⁡θij)/2z = (1-\cos\theta_{ij})/2 with θij\theta_{ij} the opening angle between the momentum vectors of the two objects, QTQ_{\mathrm{T}} is the sum of the transverse energies of all objects included in the sum, and σ\sigma is the integrated cross section used for normalization, with dσ\mathrm{d}\sigma denoting the corresponding differential cross-section element.

Many measurements of this observable were performed in which the input objects are hadronic jets produced in electron-positron [122, 123, 124, 125, 126] and proton–proton collisions [127, 128, 129]. Recently, measurements using the charged constituents within hadronic jets drove interest in these observables by accessing a highly collinear regime which is sensitive to the transition between perturbative and non-perturbative QCD [130]. Measurements of the two-point energy correlator using the charged particles within jets were performed in electron-positron [131, 132], proton–proton [120, 133], and Pb-Pb collisions [134, 135].

The EEC corresponds to an energy-weighted statistical moment of the pair distribution as a function of the opening angle. Unfolding the EEC with conventional binned techniques requires significant approximations because the observable is constructed from pairwise combinations of particles after detector reconstruction. In contrast, the published event-level representation allows the EEC to be constructed directly from the unfolded charged-particle kinematics, avoiding such approximations. This paper utilizes the fully differential charged-particle phase space measurement to measure the EEC with all final state charged particles without such approximations. A further advantage of using all charged particles in the event, rather than only those assigned to reconstructed jets, is that the measurement extends over the full angular range of zz without truncation at the jet radius. This is the first measurement of the EEC using all charged particles in the final state in proton–proton collisions.

The measurement of the EEC observable as a function of the angular separation observable zz is shown in Figure 5. The measurement achieves a precision of approximately 5% or better over the full range of zz. The fractional uncertainties and covariance matrix associated with this measurement can be reproduced using the published event representations and analysis code [65, 66].

Figure 5. Measurement of the two-point energy correlator (EEC), defined in Eq. (2), as a function of the angular-separation variable zz. Uncertainties on the measurement are shown as error bars and uncertainties on the generator predictions are shown as shaded bands.

Three distinct scaling regimes are evident in the behavior of the EEC as a function of zz. At large angles (z>10−2z > 10^{-2}), a linear evolution of the EEC with zz is observed, corresponding to non-interacting particles. Radiation at these wide angles is dominated by underlying event processes which produce no coherent energy flow. At intermediate angles, between z∼10−3z \sim10^{-3} and z∼10−4z \sim10^{-4}, the EEC is constant with zz. This deviation from the linear scaling behavior expected from a uniform radiation pattern is produced by the QCD dynamics of interacting quarks and gluons. At small angles (z<10−4z < 10^{-4}), the linear scaling behavior is restored. This is evidence of the confinement of quarks and gluons into hadrons which do not interact and produce no coherent energy flow. Significant discrepancies between the data and both generator predictions are observed in the smallest-zz region, most notably for the SHERPA prediction.

Intrinsic dimensionality of jets

Intrinsic dimensionality (ID) is a property of a dataset that quantifies the dimensionality of the low-dimensional manifold on which the data approximately lie within a high-dimensional feature space [136]. The manifold hypothesis suggests that this dimensionality can be substantially lower than the dimensionality of the feature space naively used to describe data. Recent theoretical work has suggested using ID estimators on jet-substructure data as useful observables for studying QCD and for testing BSM theories [137, 138]. However, no ID estimator has ever been measured in particle physics data, given they are properties of a data set rather than a single jet or event, making it impossible to construct a response matrix. Unbinned unfolding methods circumvent this problem, since they do not require a response matrix. This, combined with the full charged-particle phase space nature of this measurement, makes a measurement of a physically meaningful ID estimator possible for the first time. This paper focuses on the nearest neighbor intrinsic dimension (NNID) estimator [139], which was explored in Ref. [138] for tests of BSM theories, to study the ID of a selection of jets taken from the full charged-particle phase space measurement.

Any ID estimator requires a suitable distance metric, a function for calculating the distance between two points that satisfies the triangle inequality. Reference [137] introduced the energy movers distance (EMD) as a suitable distance metric for jet data. This metric quantifies the minimal “cost” required to rearrange the radiation pattern of one jet into that of another jet. The per-constituent energy weighting used in the calculation of the EMD ensures that the metric is infrared and collinear safe, meaning it is insensitive to the addition of low momentum or collinear jet constituents. The first step in measuring the NNID is to calculate the EMD between each pair of jets in the sample. Jets, clustered with the anti-ktk_{\mathrm{t}} algorithm [140] with a radius parameter of R=1.0R=1.0 and required to satisfy ∣η∣<2.5|\eta|<2.5 and pT∈[330,370]p_{\mathrm{T}} \in[330,370] GeV, are selected from the reweighted particle-level simulated events used to build the fully differential charged-particle phase space measurement. The narrow transverse momentum interval allows the intrinsic dimensionality to be studied at an approximately fixed jet energy scale. Only the leading jet is taken from each event and is assigned the weight of the event. Because the computational cost of evaluating the EMD for all jet pairs scales quadratically with the number of jets, the sample is randomly reduced to 50 000 jets for this study.

Calculation of the NNID proceeds by calculating the ratio of the EMDs between each jet and its iith and jjth nearest neighbor, with nearest neighbors determined by the EMD. Then a minimization of a likelihood function defined over these ratios is performed to extract the NNID. This study follows the recommendations of Ref. [138] and fixes j=2ij = 2i, making the only free parameter the choice of the algorithmic parameter ii. The NNID is calculated for logarithmically spaced choices of ii. Each of these choices represents a different observable, while the jets used to calculate the observable remain fixed. To plot the NNID as a function of a physically meaningful quantity, the median EMD between a jet and its iith nearest neighbor is also calculated for each choice of ii. This procedure is repeated for each weight variation to produce uncertainties on the NNID and the median EMD. The NNID is then plotted against the median EMD in Figure 6.

Figure 6. Measurement of the nearest neighbor intrinsic dimension (NNID) as a function of the median EMD between jets used to estimate the NNID. The measurement uses anti-ktk_t jets with R=1.0R = 1.0, restricted to pTjet∈[330,370]p_T^{\mathrm{jet}} \in[330, 370] GeV. Shaded bands correspond to the total uncertainty on the NNID measurement and generator predictions. Figure 6(a) shows the NNID at low values of the median EMD, and Figure 6(b) shows the NNID at high values.

For both the measurement and theory predictions, the NNID decreases as a function of the median EMD. This is an expected property of intrinsic dimension estimators. When considering jets that have similar substructure (and thus low median EMD), there are many dimensions required to describe how the fine details of each jet’s substructure differs from the other. As the EMD between the jets used to estimate the NNID increases, the variance in the jet data becomes confined to a few directions in phase space and the intrinsic dimensionality decreases. The NNID decreases approximately linearly with the median EMD at values of the median EMD below 16 GeV, and becomes approximately constant for values of the median EMD above 30 GeV. Between these regimes there is a minimum in the NNID, first decreasing then slightly increasing with larger median EMD. The theoretical predictions provided by MG5+Py8 and Sherpa generally describe the data within uncertainties, but some departure between the data and generator level predictions is visible, especially for MG5+Py8 in the region of the minimum. The decrease and subsequent increase in the NNID at intermediate scales is present in both generators, but is more pronounced than what is observed in data.

Finally, across all scales the intrinsic dimensionality is fewer than ten: much smaller than the many tens of dimensions required to describe a jet in terms of the four-vectors of its constituents. This observation provides direct experimental evidence that QCD dynamics constrain physically realized jet configurations to occupy only a small subset of the available feature space. Both generators reproduce this qualitative behavior, although modest differences with the data remain at the largest distance scales.

Conclusion

This paper has presented the first published unbinned collider cross-section measurement performed directly in the full phase space of final-state charged particles. Rather than publishing a finite collection of differential cross sections, the measurement is represented by a set of particle-level events together with associated unfolded event weights, allowing arbitrary observables to be constructed after publication. Realizing this measurement required extending Omnifold to operate on high-dimensional variable-length event representations using transformer neural networks together with large-scale pretraining on simulated data. The trained neural networks produced in the Omnifold procedure are also released publicly.

Four representative applications were presented, spanning event-level observables, jet-shape measurements, energy correlators, and intrinsic-dimensionality estimators. Together these examples demonstrate that a single published event-level representation can reproduce conventional unfolded measurements, support precision studies of observables that would otherwise require dedicated detector calibrations or difficult unfolding procedures, and enable entirely new classes of measurements that are inaccessible to traditional response-matrix techniques.

These four applications illustrate only a small subset of the possible uses of the published measurement. To facilitate future studies, the particle-level event sample and corresponding unfolded event weights are released publicly [65]. Example analysis code is also provided, reproducing each study presented in this paper and illustrating recommended workflows for downstream analyses [66]. Producing this measurement required fine-tuning many thousands of neural networks within the iterative Omnifold procedure, representing a significant computational undertaking. Approximately 400 000 NVIDIA A100 GPU hours were required to produce the final results. Continued advances in high-performance computing resources, together with methodological improvements that reduce the computational cost of uncertainty propagation, will make similar measurements increasingly practical in future analyses.

The techniques presented here provide a path toward treating unfolded event-level datasets as long-lived scientific data products, extending the impact of collider measurements well beyond the specific observables envisioned at the time of publication. Event-level unfolded measurements of this type are expected to become an increasingly important component of the LHC physics program, complementing traditional differential cross-section measurements while substantially expanding their long-term scientific utility.

Acknowledgements

We thank CERN for the very successful operation of the LHC and its injectors, as well as the support staff at CERN and at our institutions worldwide without whom ATLAS could not be operated efficiently.

The crucial computing support from all WLCG partners is acknowledged gratefully, in particular from CERN, the ATLAS Tier-1 facilities at TRIUMF/SFU (Canada), NDGF (Denmark, Norway, Sweden), CC-IN2P3 (France), KIT/GridKA (Germany), INFN-CNAF (Italy), NL-T1 (Netherlands), PIC (Spain), RAL (UK) and BNL (USA), the Tier-2 facilities worldwide and large non-WLCG resource providers. Major contributors of computing resources are listed in Ref. [141].

We gratefully acknowledge the support of ANPCyT, Argentina; YerPhI, Armenia; ARC, Australia; BMWFW and FWF, Austria; ANAS, Azerbaijan; CNPq and FAPESP, Brazil; NSERC, NRC and CFI, Canada; CERN; ANID, Chile; CAS, MOST, NSFC and FRFCU, China; Minciencias, Colombia; MEYS CR, Czech Republic; DNRF and DNSRC, Denmark; IN2P3-CNRS and CEA-DRF/IRFU, France; SRNSFG, Georgia; BMFTR, HGF and MPG, Germany; GSRI, Greece; RGC and Hong Kong SAR, China; ICHEP and Academy of Sciences and Humanities, Israel; INFN, Italy; MEXT and JSPS, Japan; CNRST, Morocco; NWO, Netherlands; RCN, Norway; MNiSW, Poland; FCT, Portugal; MNE/IFA, Romania; MSTDI, Serbia; MSSR, Slovakia; ARIS and MVZI, Slovenia; DSI/NRF, South Africa; MICIU/AEI, Spain; SRC and Wallenberg Foundation, Sweden; SERI, SNSF and Cantons of Bern and Geneva, Switzerland; NSTC, Taipei; TENMAK, Türkiye; STFC/UKRI, United Kingdom; DOE and NSF, United States of America.

Individual groups and members have received support from BCKDF, CANARIE, CRC and DRAC, Canada; CERN-CZ and PRIMUS, Czech Republic; COST, ERC, ERDF, Horizon 2020 and Marie Skłodowska-Curie Actions, European Union; Investissements d’Avenir Labex, Investissements d’Avenir Idex and ANR, France; DFG and AvH Foundation, Germany; Herakleitos, Thales and Aristeia programmes co-financed by EU-ESF and the Greek NSRF, Greece; BSF-NSF and MINERVA, Israel; NCN and NAWA, Poland; La Caixa Banking Foundation, CERCA and AGAUR programs from Generalitat de Catalunya and PROMETEO and GenT Programmes Generalitat Valenciana, Spain; Göran Gustafssons Stiftelse, Sweden; The Royal Society and Leverhulme Trust, United Kingdom; Eric and Wendy Schmidt Fund for Strategic Innovation, United States of America.

In addition, individual members wish to acknowledge support from Chile: Agencia Nacional de Investigación y Desarrollo (FONDECYT 1230812, FONDECYT 1230987, FONDECYT 1240721, FONDECYT 1240864, FONDECYT 3240661); China: Fundamental Research Funds for the Central Universities (010-63263105), Chinese Ministry of Science and Technology (MOST-2023YFA1605700, MOST-2023YFA1609300), National Natural Science Foundation of China (NSFC 12275265, NSFC-W2543005); Czech Republic: Czech Science Foundation (GACR - 24-11373S), Ministry of Education Youth and Sports (ERC-CZ-LL2327, FORTE CZ.02.01.01/00/22_008/0004632); EU: H2020 European Research Council (ERC - 101002463); European Union: European Research Council (BARD No. 101116429, ERC - 101219398, ERC 101089007), European Regional Development Fund (HE COFUND GA No. 101081355, ERDF), Marie Sklodowska-Curie Actions (GAP-101168829); France: Agence Nationale de la Recherche (ANR-22-EDIR-0002, ANR-24-CE31-0504-01); Germany: Deutsche Forschungsgemeinschaft (DFG - 469666862); China: Research Grants Council (GRF); Italy: Istituto Nazionale di Fisica Nucleare (LHC-MIUR - 28003/2025), Ministero dell’Università e della Ricerca (NextGenEU I53D23000820006 M4C2.1.1, SOE2024_0000023); Japan: Japan Society for the Promotion of Science (JSPS KAKENHI JP25H0063, JSPS KAKENHI JP22H04944, JSPS KAKENHI JP24K23939, JSPS KAKENHI JP24KK0251, JSPS KAKENHI JP25H00650, JSPS KAKENHI JP25H01291, JSPS KAKENHI JP25K01011, JSPS KAKENHI JP25K01023, JSPS KAKENHI JP25KK0047); Poland: Polish National Science Centre (NCN OPUS 2023/51/B/ST2/02507, NCN OPUS nr 2022/47/B/ST2/03059, UMO-2019/34/E/ST2/00393, UMO-2021/42/E/ST2/00350, UMO-2022/47/O/ST2/00148, UMO-2023/49/B/ST2/04085, UMO-2023/51/B/ST2/00920, UMO-2024/53/N/ST2/00869); Spain: Agència de Gestió d’Ajuts Universitaris i de Recerca. (AGAUR - 2023 BP 00141), Ministry of Science and Innovation (RYC2019-028510-I, RYC2020-030254-I, RYC2021-031273-I, RYC2022-038164-I), Ministerio de Ciencia, Innovación y Universidades/Agencia Estatal de Investigación (EU NextGenerationEU (PRTR-C17.I1), PID2022-142604OB-C22); Sweden: Carl Trygger Foundation (Carl Trygger Foundation CTS 22:2312), Swedish Research Council (Swedish Research Council 2023-04654, VR 2021-03651, VR 2022-03845, VR 2022-04683, VR 2023-03403, VR 2024-05451, VR 2025-05940), Knut and Alice Wallenberg Foundation (KAW 2023.0366); United Kingdom: The Binks Trust, Royal Society (NIF-R1-231091); United States of America: U.S. Department of Energy (ECA DE-AC02-76SF00515), John Templeton Foundation (John Templeton Foundation 63206), Neubauer Family Foundation.

Appendix

Usage recommendations for the unbinned measurement

The primary result of this paper is a published event-level representation of the measured particle-level cross section, consisting of particle-level MC events together with their associated unfolded weights. This data product is released publicly to enable the construction of new observables and analyses beyond those presented in this paper. The recommendations below summarize the conditions under which the measurement was validated and describe best practices for downstream analyses. The same recommendations are implemented in the accompanying public codebase and are reproduced here for convenience.

1. Validation using the pseudo-measurement Any analysis should always be checked using the pseudo-measurement against the target, to ensure a reliable result is obtained before drawing any conclusion based on the measurement with real data. For any binned measurement, the binning should be chosen such that each bin contains sufficient statistics and the unfolded pseudomeasurement shows good statistical agreement with the target when tested using pseudodata. Examples of such tests are performed in each of the example notebooks before the results of the data measurement are shown. The key step of this procedure is to construct the measurement covariance matrix that encodes the correlation between bins. The χ2\chi^2 statistic and associated pp-value are then obtained with a few lines of code that performs matrix operations. The outcome of these tests will depend on the assumed nuisance parameter smoothing and correlation model as discussed in the next bullet.

2. Minimum jet transverse momentum When studying jet observables, jets should satisfy pT>5p_{\mathrm{T}} > 5 GeV. For observables involving the leading jet, such as ΔR(ℓℓ,j1)\Delta R(\ell\ell, j_1), this requirement should be applied to the leading jet. For dijet observables such as mjjm_{jj}, it should be applied to both jets.

3. Sufficient Monte Carlo statistical precision Each bin should contain a minimum of 5000 effective events defined by neff≡(∑iwi)2∑iwi2n_{\mathrm{eff}} \equiv\frac{(\sum_i w_i)^2}{\sum_i w_i^2}. This requirement ensures that the Monte Carlo statistical uncertainty is sufficiently small for stable uncertainty estimates.

4. Sufficient data statistical precision The statistical uncertainty from the data should not exceed 15% in any bin, ensuring that the measurement is supported by sufficient data statistics. This criterion is most likely fulfilled when neff>5000n_{\mathrm{eff}} > 5000.

5. Uncertainty smoothing and correlation model The default recommendation is to smooth the hidden-variable uncertainties using the Gaussian-kernel procedure provided in utils.py and to treat each systematic uncertainty source as fully correlated across bins. However, there are cases where adjustments should be made to this model. For example, in the default results presented in standard_obs_data_results.ipynb the hidden variable uncertainties are not smoothed for the mj1m_{j1} and mj2m_{j2} observables. This is because these two-point systematic uncertainties are large and have a strong shape, meaning smoothing them would bias the uncertainty and result in poor pp-values.

Uncertainties with strong shape may additionally switch in sign in the middle of the distribution, potentially resulting in poor pp-values due to the assumption that bins are fully correlated. This is one example where other correlation models may be explored. Since the true correlation model is not known, any alternative correlation model should first be validated using the pseudomeasurement before being applied to the data. For instance, if four bins are used, the systematic uncertainty may be split into four components (‘nuisance parameters’), one per bin. More sophisticated decorrelation schemes could be explored for which each uncertainty component has a smooth uncertainty amplitude spanning the full range, such that at any point (i.e., for any bin) this uncertainty adds in quadrature to match the total uncertainty of the two-point systematics.

Drell–Yan-only theory comparisons

The public theory prediction files contain a vector named sample, whose integer values identify simulated events originating from the Drell–Yan (0), diboson (1), and electroweak (2) processes described in § 2. These files can therefore be used to extract predictions for each individual component. This is useful, for example, when comparing the measurement with a novel Drell–Yan prediction while no corresponding predictions are available for the electroweak ZjjZjj or ZVZV contributions. In this case, the public electroweak and diboson predictions can be added to the user-provided Drell–Yan prediction to construct the full prediction for comparison with the published measurement. The published measurement is agnostic to the underlying production process of the particle-level events which are used to construct the published measurement and therefore does not support selections based on the process labels. An example illustrating how to extract these components and their associated uncertainties for the pTμμp_{\mathrm{T}}^{\mu\mu} distribution is provided in Drell_Yan_results.ipynb.

Not every observable that can be constructed from the published particle-level events is necessarily constrained by the measurement. For example, the public event representation includes the particle masses, allowing users to identify the hadron species of individual particles. However, the relative abundances of different hadron species are not directly constrained by the detector-level information used in the unfolding. Similarly, the published measurement includes event-level process labels (Drell–Yan, electroweak ZjjZjj, and diboson), which identify the simulated production process from which each event originated. These labels are retained for technical consistency with the public theory prediction files, but are not constrained by the unfolding and should therefore not be used to define measured subsamples or fiducial cross sections. Consequently, observables that depend strongly on the particle-level hadron composition or on the event process labels remain prior dependent and will generally fail to satisfy the validation procedure described above. All downstream analyses should therefore follow the recommendations above, with particular emphasis on validation using the pseudodata measurement before interpreting results obtained from the data.

In almost all cases, a successful validation using pseudodata demonstrates that the unfolding procedure is capable of reproducing the target cross section in the bins that the user has defined. It is, however, possible that the predictions of the MG5+PY8 and SHERPA 2.2.11 generators happen to agree for some observable that is not constrained by the unfolding. In this case the pseudodata validation is trivial, and the measurement should not be considered capable of constraining this particular observable. When using the information that is known to be unconstrained by the unfolding, the user should verify that the pseudodata validation is non-trivial before proceeding.

The validation procedure described above assumes that the downstream analysis projects the event-level representation to construct a binned measurement. In this case, an appropriate goodness-of-fit test is the binned χ2\chi^2 procedure implemented in the example notebooks. More advanced applications may instead operate directly on the high-dimensional event representation without introducing intermediate histograms. Defining statistically rigorous goodness-of-fit tests for such fully unbinned analyses remains an open area of research. Consequently, no recommended procedure is currently provided in the public codebase. Users pursuing such applications should develop an appropriate validation strategy using the pseudodata measurement before drawing conclusions from the corresponding data analysis.

Details of pseudodata measurement and statistical tests

The primary validation of this measurement is based on a pseudodata study, in which simulated events are modified such that their detector-level distributions closely resemble those observed in data while retaining a known particle-level target. Because the particle-level target remains known, the complete measurement chain—including background subtraction, unfolding, uncertainty propagation, and statistical interpretation—can be tested under realistic conditions. In contrast, this validation cannot be performed using collision data because the corresponding particle-level distribution is unknown.

The pseudodata are constructed by combining independent samples of the Sherpa Drell–Yan, electroweak, diboson, and Powheg v2 ttˉt\bar{t} processes. Smooth reweighting functions are derived from detector-level observables and applied at particle level such that, after detector simulation, the resulting detector-level distributions closely reproduce those observed in data. The reweighted sample is subsequently divided into statistically independent detector- and particle-level samples. The detector-level sample is randomly downsampled to match the effective statistical precision of the collision data and serves as the pseudodata input to the full analysis chain. The particle-level sample retains the full Monte Carlo statistics and provides a known target against which the unfolded measurement is compared. This combination yields a realistic validation of the measurement while minimizing the statistical uncertainty on the reference target.

All uncertainties are propagated independently for the pseudodata-based and data-based measurements. The resulting pseudodata-based measurement is then compared with the known particle-level target to assess the statistical consistency of the measurement. For most applications of this measurement, this should be done with a binned χ2\chi^2 statistical test, performed as follows:

  1. The nominal measurement and all associated weight variations are projected onto the observable of interest. The varied histograms should then be used to construct a covariance matrix. The covariance matrix encodes both the variance of each bin and the correlations between bins arising from shared systematic uncertainties and statistical fluctuations. It therefore represents the complete uncertainty model of the binned measurement. Detailed illustration of how to perform these steps is provided in the public codebase.

  1. The particle-level pseudodata sample should be binned in the same observable and bins as the pseudodata measurement.

  1. A χ2\chi^2 test statistic is then calculated as

χ2=(m−t)TΣ−1(m−t),(3)\chi^2 = (\mathbf{m}-\mathbf{t})^T \Sigma^{-1} (\mathbf{m}-\mathbf{t}), \tag*{(3)}

where m\mathbf{m} is the vector of measured bin contents, t\mathbf{t} is the corresponding particle-level pseudodata target, and Σ\Sigma is the covariance matrix. The test statistic quantifies the statistical compatibility between the measurement and the known target while accounting for the statistical and systematic uncertainties, including their correlations between bins.

  1. The resulting χ2\chi^2 test statistic should be compared to a χ2\chi^2 distribution with the number of degrees of freedom ndofn_{\mathrm{dof}} equal to the number of bins in the histogram, and a pp-value should be calculated. A sufficiently large pp-value indicates that the measurement is statistically consistent with the known target under the assumed uncertainty model. The target hypothesis is not rejected at the 5% significance level when the pp-value is at least 0.05.

The results of these statistical tests for the nchn_{\mathrm{ch}} and HTH_{\mathrm{T}} observables, for which the real data measurements are presented in § 5.1, are presented in Figure 7. Here the pseudomeasurement is compared with the known particle-level pseudodata target (the nchn_{\mathrm{ch}} distribution is shown in Figure 2), the pseudodata uncertainty breakdown, and the corresponding correlation matrices. The uncertainties on the pseudomeasurement are generally very similar in magnitude and shape to the uncertainties on the corresponding real measurement. A numerical summary of the corresponding tests for all 26 standard event-level observables is given in Table 1. The observed pp-values indicate that the unfolded pseudodata measurements are statistically consistent with their known particle-level targets for all representative observables considered. These tests provide evidence that the assigned uncertainties adequately cover the residual biases of the unfolding procedure for the chosen observables and binnings.

Observablendofn_{\mathrm{dof}}χ2\chi^2pp
HTH_{\mathrm{T}}66.210.400
nchn_{\mathrm{ch}}52.410.790
nch,j1n_{\mathrm{ch},j1}610.40.110
nch,j2n_{\mathrm{ch},j2}54.720.450
ημ1\eta_{\mu1}1411.10.678
ημ2\eta_{\mu2}144.340.993
mj1m_{j1}615.60.016
mj2m_{j2}49.230.056
pT,μ1p_{\mathrm{T},\mu1}64.570.600
pT,μ2p_{\mathrm{T},\mu2}68.200.224
pTμμp_{\mathrm{T}}^{\mu\mu}54.430.490
pT,j1p_{\mathrm{T},j1}611.30.081
pT,j2p_{\mathrm{T},j2}42.120.713
ϕμ1\phi_{\mu1}1613.50.634
ϕμ2\phi_{\mu2}1617.30.364
ϕj1\phi_{j1}167.020.973
ϕj2\phi_{j2}167.770.955
τ1,j1\tau_{1,j1}73.320.854
τ1,j2\tau_{1,j2}56.970.223
τ2,j1\tau_{2,j1}711.90.105
τ2,j2\tau_{2,j2}55.400.369
τ3,j1\tau_{3,j1}41.870.760
τ3,j2\tau_{3,j2}41.920.750
yμμy_{\mu\mu}1415.70.334
yj1y_{j1}1814.00.727
yj2y_{j2}105.690.840

Table 1. χ2\chi^2 test statistics and associated pp-values for the Omnifold pseudodata-based measurements. A pp-value of at least 0.05 indicates that the null hypothesis of agreement between the unfolded measurement and the particle-level pseudodata target is not rejected at the 95% confidence level. In these tests, Gaussian kernel smoothing is applied to the hidden-variable unfolding uncertainty for all observables except mj1m_{j1} and mj2m_{j2}. A minimum jet pTp_{\mathrm{T}} requirement of pT,j1>5p_{\mathrm{T},j1} > 5 GeV for leading-jet observables and pT,j2>5p_{\mathrm{T},j2} > 5 GeV for subleading-jet observables is imposed. No minimum jet requirement is applied to observables with no explicit dependence on hadronic activity.

Figure 7. Validation of the pseudodata measurement for the HTH_{\mathrm{T}} and nchn_{\mathrm{ch}} observables. (a) Comparison of the pseudodata measurement of the HTH_{\mathrm{T}} observable with the known particle-level pseudodata target. The analogous plot for the nchn_{\mathrm{ch}} observable is shown in Figure 2(a). Breakdowns of the various pseudodata measurement uncertainties are shown for (b) HTH_{\mathrm{T}} and (d) nchn_{\mathrm{ch}}. The corresponding uncertainty correlation matrices are shown for (c) HTH_{\mathrm{T}} and (e) nchn_{\mathrm{ch}}.

A key advantage of the published event-level representation is that observables can be rebinned after publication without repeating the unfolding. To illustrate this flexibility, Table 2 presents the results of statistical comparisons between the unfolded distributions and the pseudodata targets for several binning choices of the nchn_{\mathrm{ch}} and HTH_{\mathrm{T}} observables. For HTH_{\mathrm{T}}, good agreement with the particle-level target is observed for all three binnings considered. For nchn_{\mathrm{ch}}, the pp-value decreases as the binning becomes finer, reaching p=0.041p = 0.041 for the finest binning. This behavior demonstrates the importance of validating any new binning with the accompanying pseudodata measurement before interpreting results obtained from collision data.

ObservableNeffN_{\mathrm{eff}} / binndofn_{\mathrm{dof}}χ2\chi^2pp
nchn_{\mathrm{ch}}30,00065.010.543
HTH_{\mathrm{T}}30,00063.420.755
nchn_{\mathrm{ch}}20,0001013.20.214
HTH_{\mathrm{T}}20,000105.960.819
nchn_{\mathrm{ch}}10,0002032.20.041
HTH_{\mathrm{T}}10,0002014.10.826

Table 2. Results of the χ2\chi^2 goodness-of-fit tests for several binning choices of the nchn_{\mathrm{ch}} and HTH_{\mathrm{T}} observables. The bin edges are chosen such that each bin contains approximately NeffN_{\mathrm{eff}} effective events. A pp-value of at least 0.05 indicates that the null hypothesis of agreement between the unfolded measurement and the particle-level pseudodata target is not rejected at the 95% confidence level. These results illustrate the importance of validating any new binning choice using the accompanying pseudodata measurement. Gaussian kernel smoothing is applied to the hidden-variable uncertainty.

For the finest nchn_{\mathrm{ch}} binning, the reduced pp-value is traced to the current treatment of the two-point unfolding uncertainty. Although the assigned uncertainty provides sufficient overall coverage, the assumed correlation model does not fully capture the pronounced shape of the corresponding method bias. As discussed in the user recommendations in Appendix § 8.1, alternative smoothing or decorrelation schemes may be explored when appropriately motivated, but any such choice should first be validated using the pseudodata measurement.

As discussed in § 4.3, an additional pseudodata validation was performed to probe regions in which the alternative and nominal Drell–Yan generators, SHERPA and MG5+Py8, respectively, differ strongly. A ParT classifier, identical to that described in § 3.3, is used to distinguish the alternative pseudodata sample, whose Drell–Yan component is generated with SHERPA, from the nominal sample, whose Drell–Yan component is generated with MG5+Py8. This task is identical to the pretraining task described in § 3.4, and one of the networks produced during pretraining is therefore repurposed for this study.

The high-SHERPA-probability tail of the classifier output identifies a region of relatively poor support from the nominal sample. In this region, the density of nominal events is approximately an order of magnitude smaller than that of the SHERPA-based pseudodata target, providing a stringent test of the unfolding procedure. The agreement between the unfolded results and the known particle-level target is evaluated in six bins concentrated in this region of the classifier output in logit space, with each bin containing approximately 5000 effective events in the reweighted nominal sample, following the usage recommendations in Appendix § 8.1.

The results are shown in Figure 8. Despite the substantially lower density of nominal events in this region, the pseudomeasurement is consistent with the known particle-level pseudodata target, with χ2/ndof=0.424\chi^2 / n_{\mathrm{dof}} = 0.424 and a pp-value of 0.863.

Figure 8. Results of the validation test using pseudodata performed in six bins concentrated in the high-Sherpa-probability tail of the classifier output in logit space. The nominal distribution defining the prior used in the unfolding is shown as triangles, the known particle-level pseudodata target is shown as open circles, and the result of the pseudomeasurement is shown as points. Despite the substantially lower density of nominal events in this region, good agreement is observed, with a p-value of 0.863.

Comparison to IBU and Multifold measurements

This appendix compares the Omnifold measurements of a series of standard event-level observables with analogous results obtained using IBU and the previous Multifold measurement [4]. All three analyses use identical event selections and target the same fiducial phase space, enabling a direct comparison of the measured cross sections.

The uncertainty models are largely identical. Since IBU does not rely on neural-network training, the corresponding measurements do not include a training stochasticity uncertainty. Aside from this difference, the IBU and Omnifold measurements propagate the same uncertainty sources. The Multifold analysis differs slightly in the treatment of several subdominant uncertainty sources.

Measurement comparison

Figure 9–Figure 13 present comparisons of the three measurements for the 24 previously measured observables and the two additional observables considered here. Excellent agreement is observed for all observables considered. Although the projected measurements are similar, the scope of the underlying measurements differs substantially.

Figure 9. Comparison of the cross-section measurements obtained with IBU, Multifold [4] and Omnifold (labeled ‘This result’). The panels show: the charged-particle multiplicity of the (a) leading and (b) subleading charged-hadron jet; the pseudorapidity of the (c) leading and (d) subleading muon; the mass of the (e) leading and (f) subleading charged-hadron jet. Excellent agreement is observed among the three measurements for all observables, despite the substantially different scope of the underlying measurements.

Figure 13. Comparison of the cross-section measurements obtained with IBU (labeled ‘IBU’) and OMNIFOLD (labeled ‘This result’) for the (a) nchn_{\mathrm{ch}} and (b) HTH_{\mathrm{T}} observables. These observables were not measured using MULTIFOLD [4].

Each IBU result corresponds to an independent unfolding performed for a single observable. Consequently, measuring an additional observable requires repeating the complete unfolding procedure. In contrast, Multifold simultaneously unfolds 24 observables within a single event-level measurement, although it is not differential in the nchn_{\mathrm{ch}} and HTH_{\mathrm{T}} observables and therefore no corresponding Multifold results are shown in Figure 13. Finally, the Omnifold measurement is differential in the full kinematics of the final-state charged particles. The same published event-level representation is used to construct all 26 observables shown here, and the four example analyses presented in § 5.

Figure 10. Comparison of the cross-section measurements obtained with IBU, Multifold [4] and Omnifold (labeled ‘This result’) for the transverse momentum of the (a) leading and (b) subleading muon, the transverse momentum of the (c) leading and (d) subleading charged-hadron jets, and (e) the rapidity of the dimuon system.

Figure 11. Comparison of the cross-section measurements obtained with IBU, Multifold [4] and Omnifold (labeled ‘This result’) for the azimuthal angle of the (a) leading and (b) subleading muons, the (c) leading and (d) subleading charged-hadron jets, and the 1-subjettiness [142] τ1\tau_1 of the (e) leading and (f) subleading charged-hadron jets.

Figure 12. Comparison of the cross-section measurements obtained with IBU, Multifold [4] and Omnifold (labeled ‘This result’) for the 2-subjettiness τ2\tau_2 of the (a) leading and (b) subleading charged-hadron jets, the 3-subjettiness τ3\tau_3 of the (c) leading and (d) subleading charged-hadron jets, and the rapidity of the (e) leading and (f) subleading charged-hadron jets.

Total uncertainty comparison

Figure 14–Figure 18 compare the total uncertainties obtained with the Omnifold, Multifold, and IBU measurements for each of the 26 observables. Overall, the projected uncertainties are remarkably similar across the three approaches. Despite deriving a measurement that is differential in the full charged-particle phase space, the Omnifold measurement incurs little or no uncertainty penalty relative to dedicated one-dimensional IBU measurements for most observables. For some jet observables, most notably the leading and subleading jet masses, the Omnifold uncertainties are even smaller than those obtained with IBU. The largest improvements relative to the previous Multifold measurement are observed for observables dominated by the muon kinematics, reflecting the improved stability of the neural-network training used in the present work.

Figure 14. Comparison of the total uncertainty estimates obtained with IBU, Multifold [4] and Omnifold (labeled ‘This result’). The panels show the charged-particle multiplicity of the (a) leading and (b) subleading charged-hadron jet, (c) the leading-muon pseudorapidity, (d) the subleading-muon pseudorapidity, the mass of the (e) leading and (f) subleading charged-hadron jet.

Figure 18. Comparison of the total uncertainty estimates obtained with IBU (labeled ‘IBU’) and Omnifold (labeled ‘This result’) for the (a) nchn_{\mathrm{ch}} and (b) HTH_{\mathrm{T}} observables. These observables were not measured using Multifold [4].

Figure 15. Comparison of the total uncertainty estimates obtained with IBU, Multifold [4] and Omnifold (labeled ‘This result’) for the pTp_{\mathrm{T}} of the (a) leading and (b) subleading muon, the (c) transverse momentum and the (d) rapidity of the dimuon system, and the transverse momentum of the (e) leading and (f) subleading charge-hadron jets.

Figure 16. Comparison of the total uncertainty estimates obtained with IBU, Multifold [4] and Omnifold (labeled ‘This result’) for the azimuthal angle of the (a) leading and (b) subleading muon, and (c) leading and (d) subleading charged-hadron jets, and the 1-subjettiness [142] τ1\tau_1 of the (e) leading and (f) subleading charged-hadron jets.

Figure 17. Comparison of the total uncertainty estimates obtained with IBU, Multifold [4] and Omnifold (labeled ‘This result’) for the 2-subjettiness τ2\tau_2 of the (a) leading and (b) subleading charged-hadron jets, the 3-subjettiness τ3\tau_3 of the (c) leading and (d) subleading charged-hadron jets, and the rapidity of the (e) leading and (f) subleading charged-hadron jets.

Hidden variable uncertainty comparison

Figure 19 and Figure 20 compare the hidden-variable uncertainty estimated for the IBU, MULTIFOLD [4], and OMNIFOLD measurements. The figures show the signed uncertainty as opposed to the magnitude as in Appendix § 10.§ 2, since a single uncertainty is shown rather than many uncertainties added in quadrature. The dominant hidden variables in this analysis, most notably the particle-level hadron composition, are only weakly constrained by the detector-level information. Consequently, even a highly differential unfolding cannot substantially reduce this uncertainty source, and comparable hidden-variable uncertainties are expected for all three approaches, as observed in Figure 19 and Figure 20. In some observables, such as the subleading charged-hadron jet mass (Figure 19(b)), the OMNIFOLD measurement yields a smaller hidden-variable uncertainty than either MULTIFOLD or IBU, whereas in others, such as pTμp_T^{\mu} (Figure 19(e)), the uncertainty is slightly larger. These larger uncertainties are primarily observed in low-statistics tails of distributions, where limited training data makes accurate estimation of the likelihood ratios used by the neural-network classifier more challenging. Nevertheless, these differences are small compared with the total uncertainty budget and have little impact on the overall precision of the projected measurements, as shown in Appendix § 10.§ 2.

Figure 19. Comparison of the hidden variable uncertainty obtained with IBU, Multifold [4] and Omnifold (labeled ‘This result’) for the masses of the (a) leading and (b) subleading charged-hadron jet, the transverse momentum of the (c) leading and (d) subleading muon, and the (e) transverse momentum and the (f) rapidity of the dimuon system.

Figure 20. Comparison of the hidden variable uncertainty obtained with IBU, Multifold [4] and Omnifold (labeled ‘This result’) for the 1 subjettiness τ1\tau_1 of the (a) leading and (b) subleading charged-hadron jet, the (c) number of charged hadrons nchn_{\mathrm{ch}}, and the (d) HTH_{\mathrm{T}}.

D Unbinned background subtraction

This appendix describes the procedure used to subtract the small ttˉt\bar{t} background from the pseudodata and data samples prior to unfolding with Omnifold. Since the unfolding is performed without binning, the background subtraction must also be performed in an unbinned manner. Following Ref. [89], the subtraction is achieved by training a neural network classifier to reweight between the data distribution, and the background-subtracted data distribution, utilizing a background model provided by a Monte Carlo generator. In the context of this analysis the unbinned subtraction is performed by training a neural network classifier to distinguish two sets of simulated events: the detector-level pseudodata sample described in § 3.5, and the same set of simulated events concatenated with a negatively weighted detector-level MC sample that describes the ttˉt\bar{t} background contribution. These negatively weighted events statistically remove the existing ttˉt\bar{t} events in pseudodata and leave only the signal contribution. Classifiers are trained to distinguish the second set from the first set, and the neural network output can then be used to reweight the pseudodata sample to match the background removed sample.

The classifiers described above are trained with simulation. Once trained the networks are evaluated over the data, and their outputs are used to remove the ttˉt\bar{t} background from the data. A dedicated uncertainty, in which the background subtraction procedure is repeated with an alternative description of the ttˉt\bar{t} background furnished by an alternative generator, accounts for the generator dependence. Both the nominal and alternative background subtractions are then propagated through the unfolding procedure to produce the final background subtraction uncertainty.

As described in § 3.1, the ttˉt\bar{t} background is small, accounting for only 0.25% of events in the selected event yield. This implies that differences between the detector-level pseudodata and the ttˉt\bar{t} subtracted pseudodata are small, making the classification task described above difficult. Initial efforts to learn the required likelihood ratio directly from the full charged-particle phase-space representation were not successful. In particular, Particle Transformer neural networks trained only on the low-level event representation described in § 3.3 were unable to isolate the ttˉt\bar{t} background contribution and remove it reliably. Instead, the background subtraction procedure is performed differentially in five event-level summary observables, selected because the background contribution is localized in the distributions of these observables. The most important of the five observables is the output of a Particle Transformer neural network, with the same architecture and optimized from scratch using the same methods as described in § 3.3 and § 3.4, that is trained to discriminate background ttˉt\bar{t} events from signal events. The background contribution to the detector-level pseudodata is localized in the tail of this highly expressive observable, making it very useful for the training of the classifiers which perform the background subtraction as described above. The four other observables used in the training are: nchn_{\mathrm{ch}}, HTH_{\mathrm{T}}, pTj1p_{\mathrm{T}}^{j1}, and mj1m^{j1}. Since the event representation used for background subtraction is fixed-length, MLP neural networks with two hidden layers of 512 nodes each and GeLU activation functions are used. As with the neural networks used in the Omnifold procedure, the MLPs were implemented and trained in PyTorch. Ensembles of 100 neural networks were trained for both the nominal and alternative background subtractions. Given that the background subtraction should only remove density, the weights produced by the networks were clipped to a maximum of 1.0, and then averaged across the ensemble to produce the final background subtraction weights.

Histograms of the background subtraction weights for the pseudodata sample are shown in Figure 21. Similar weight distributions were obtained for the data sample. The weight distributions produced using the nominal Powheg v2 +Pythia 8.230 sample and the alternative Sherpa 2.2.1 sample both have mean values close to 0.998, consistent with the expected ≈0.25%\approx0.25\% background fraction. The weight distributions however have substantially different variances, with the distribution of the Powheg v2 +Pythia 8.230 weights having a substantially larger tail. The difference between these weight distributions corresponds to the top background uncertainty which is propagated through the unfolding procedure. Illustrations of the top subtraction in two observables are shown in Figure 22. Figure 22(a) shows the ttˉ\mathrm{t\bar{t}} classifier output, where the ttˉ\mathrm{t\bar{t}} contribution is localized in the tail of the distribution. Both the nominal and alternative subtractions are able to remove the background, except in the very low-statistics tail above a classifier logit of 5.0. Figure 22(b) shows the HTH_T observable, where the background contribution is less localized but both subtractions generally succeed in removing the background.

Figure 21. Histograms of the ttˉ\mathrm{t\bar{t}} background subtraction weights produced using the nominal Powheg v2 +Pythia 8.230 sample (triangles) and the alternative Sherpa 2.2.1 sample (squares) for the background estimate. Both methods remove approximately the same overall ttˉ\mathrm{t\bar{t}} contribution (≈0.25%\approx0.25\% of the selected event yield), but assign different weights to individual events. The differences between the two sets of subtraction weights define the ttˉ\mathrm{t\bar{t}} background subtraction uncertainty.

Figure 22. The ttˉ\mathrm{t\bar{t}} background subtraction for the detector-level pseudodata sample. The pre-subtraction pseudodata is plotted as points, the signal-only pseudodata with the ttˉ\mathrm{t\bar{t}} contribution removed is plotted as circles, and the pseudodata with the nominal and alternative subtraction weights applied are shown as triangles and squares respectively. Open markers are used in the upper pad for visibility but denote the same quantity. Ratio pads show the ratio to the signal-only pseudodata, which is the target of the subtraction procedure. The subtraction is shown for the (a) ttˉ\mathrm{t\bar{t}} classifier output (logit) and (b) HTH_T observables.

Omnifold hyperparameter optimization methods and results

This appendix summarizes the hyperparameter optimization strategy used for the Omnifold procedure. Unlike conventional classification tasks, the objective of the neural networks within Omnifold is not to maximize classification accuracy but to estimate likelihood ratios as accurately as possible. Although these objectives become equivalent in the limit of infinite data and infinitely expressive models, finite training samples and limited model capacity mean that standard classification metrics, such as the area under the ROC curve, are not necessarily optimal proxies for unfolding performance. Furthermore, because Omnifold is an iterative algorithm, stable likelihood-ratio estimates that produce moderate corrections at each iteration are generally preferred over highly variable estimates that maximize classification performance in a single step.

To directly optimize the quality of the unfolded measurement, the objective function for the hyperparameter optimization was defined as the sum of the one-dimensional Wasserstein distances evaluated for the 24 observables measured in Ref. [4], together with the nchn_{\mathrm{ch}} and HTH_{\mathrm{T}} observables introduced in this work. This metric was found to correlate significantly better with the quality of the resulting reweighting than traditional classification metrics.

A complete hyperparameter optimization across every neural-network training within the Omnifold procedure would be computationally prohibitive. Hyperparameter optimization was therefore restricted to the pretraining stage and the step-1 and step-2 classifiers of the first Omnifold iteration. The chosen hyperparameters for steps one and two were then used for all subsequent Omnifold iterations. The hyperparameters that were optimized in each of the three trainings are marked in Table 3. The ranges used and final values are given in Table 4, Table 5, and Table 6. Architectural hyperparameters (e.g., the number of attention blocks and class-attention blocks) were optimized only during pretraining, since the same network architecture is subsequently reused throughout the unfolding. Training hyperparameters, including the learning-rate schedule and regularization parameters, were optimized independently for each step. All sweeps were performed using a random search method and the configuration which produced the best sum of Wasserstein distances was selected. This process was observed to be important for the overall performance of the Omnifold procedure. In particular, the learning rates, including the length of the annealing periods in the learning rate schedule, were very important to tune properly. Although the optimization required training hundreds of neural networks, the associated computational cost remained a small fraction of the total resources required for the full unfolding analysis.

HyperparameterpretrainingStep 1Step 2
Number of attention blocks✓
Number of class attention blocks✓
Attention block dropout✓✓✓
Class attention block dropout✓
MLP dropout✓
Minimum learning rate✓✓✓
Optimizer✓✓✓
Weight decay✓✓
Cosine annealing period✓✓

Table 3. Hyperparameters optimized in the pretraining, step 1, and step 2 trainings. A ✓ means the hyperparameter was optimized in that training.

HyperparameterLower boundUpper boundFinal value
Number of attention blocks386
Number of class attention blocks131
Attention block dropout0.00.30.02
Class attention block dropout0.00.50.02
MLP dropout0.00.50.0
Minimum learning rate1e-71e-31e-5
Optimizer––Lion

Table 4. Hyperparameter ranges and final values for the pretraining task. A random search with 150 runs was performed. The AdamW and Lion optimizers were considered for the optimizer choice.

HyperparameterLower boundUpper boundFinal value
Block dropout0.00.30.02
Minimum learning rate1e-71e-38e-6
Weight decay0.00.20.01
Cosine annealing period80040003000
Optimizer––Lion

Table 5. Hyperparameter ranges and final values for the step 1 task. A random search with 100 runs was performed. The AdamW and Lion optimizers were considered for the optimizer choice.

HyperparameterLower boundUpper boundFinal value
Block dropout0.00.30.02
Minimum learning rate1e-71e-31e-5
Weight decay0.00.20.01
Cosine annealing period10002000015000
Optimizer––Lion

Table 6. Hyperparameter ranges and final values for the step 2 task. A random search with 50 runs was performed. The AdamW and Lion optimizers were considered for the optimizer choice.

The ATLAS Collaboration

G. Aad 102, E. Aakvaag 17, B. Abbott 121, S. Abdelhameed 83b, K. Abeling 54, N.J. Abicht 48, S.H. Abidi 30, M. Aboelela 44, A. Aboulhorma 36e, H. Abramowicz 154, B.S. Acharya 68a,68b,m, A. Ackermann 62a, J. Ackerschott 55, C. Adam Bourdarios 4, L. Adamczyk 85a, S.V. Addepalli 146, M.J. Addison 101, J. Adelman 117, A. Adiguzel 22c, T. Adye 135, A.A. Affolder 137, Y. Afik 39, M.N. Agaras 13, A. Aggarwal 100, C. Agheorghiesei 28c, A. Ahmad 83a, F. Ahmadov 38,ad, S. Ahuja 165, X. Ai 113c, G. Aielli 75a,75b, A. Aikot 165, M. Ait Tamlihat 36e, T.P.A. Åkesson 98, D. Akiyama 170, S. Aktas 168, G.L. Alberghi 24b, J. Albert 167, U. Alberti 20, P. Albicocco 52, S. Alderweireldt 51, Z.L. Alegria 122, M. Aleksa 37, I.N. Aleksandrov 38, C. Alexa 28b, T. Alexopoulos 10, F. Alfonsi 24b, M. Algren 55, M. Alhroob 169, B. Ali 133, H.M.J. Ali 91,v, S. Ali 32, S.W. Alibocus 92, M. Aliev 34c, G. Alimonti 70a, C. Allaire 65, B.M.M. Allbrooke 149, D.R. Allen 122, J.S. Allen 101, C.S. Alley 1, E.R. Almazan 137, A. Aloisio 71a,71b, F. Alonso 90, C. Alpigiani 140, D. Alvarez Feito 37, A. Alvarez Fernandez 100, M. Alves Cardoso 55, M.G. Alviggi 71a,71b, Y. Amaral Coutinho 81b, C. Amelung 37, M. Amerl 107, T. Amezza 128, B. Amini 53, K. Amirie 158, A. Amirkhanov 38, D. Amperiadou 155, C. Anastopoulos 142, T. Andeen 11, J.K. Anders 92, A.C. Anderson 58, N. Anderson 148, A. Andreazza 70a,70b, S. Angelidakis 9, A. Angerami 41, A.V. Anisenkov 38, A. Annovi 73a, C. Antel 37, E. Antipov 148, M. Antonelli 52, F. Anulli 74a, M. Aoki 82, T. Aoki 156, M. Aparo 13, L. Aperio Bella 47, M. Apicella 31, C. Appelt 154, A. Apyan 27, R. Arakida 125, M. Arampatzi 10, S.J. Arbiol Val 86, C. Arcangeletti 52, A.T.H. Arce 50, M. Arcuri 43b,43a, J-F. Arguin 108, S. Argyropoulos 155, J.-H. Arling 47, O. Arnaez 4, A.P. Arnold 94, H. Arnold 148, G. Artoni 74a,74b, H. Asada 111, S. Asatryan 175, N.A. Asbah 37, R.A. Ashby Pickering 169, A.M. Aslam 95, J. Assahsah 36d, K. Assamagan 30, R. Astalos 29a, K.S.V. Astrand 98, S. Atashi 162, R.J. Atkin 34a, H. Atmani 36f, P.A. Atmasiddha 129, K. Augsten 133, A.D. Auriol 40, V.A. Austrup 101, A.S. Avad 94, G. Avolio 37, A. Azzam 13, D. Babal 29b, H. Bachacou 136, K. Bachas 155,p, A. Bachiu 35, E. Bachmann 49, M.J. Backes 62a, A. Badea 39, T.M. Baer 106, M. Bahmani 19, D. Bahner 53, K. Bai 124, L. Baines 94, O.K. Baker 174, D. Bakshi Gupta 8, L.E. Balabram Filho 81b, V. Balakrishnan 121, R. Balasubramanian 4, P. Balek 85a, E. Ballabene 24b,24a, F. Balli 136, L.M. Baltes 62a, W.K. Balunas 83c, I. Bamwidihi 83c, E. Banas 86, M. Bandieramonte 130, D. Banerjee 164, S. Banfi 25, L. Barc 154, M. Barakat 47, E.L. Barberio 105, D. Barberis 18b, M. Barbero 102, M.Z. Barel 116, T. Barillari 110, M-S. Barisits 37, T. Barklow 146, P. Baron 134, D.A. Baron Moreno 101, A. Baronc elli 61, A.J. Barr 127,g, J.D. Barr 96, F. Barreiro 99, J. Barreiro Guimarães da Costa 14, M.G. Barros Teixeira 131a, F. Bartels 37, R. Bartoldus 146, A.E. Barton 91, P. Bartos 29a, M. Baselga 48, A. Bassalat 65,b, M.J. Basso 159a, S. Bataju 4, R. Bate 166, R.L. Bates 58, M. Battaglia 137, D. Battulga 19, M. Bauce 74a,74b, L. Bauchhage 47, P. Bauer 25, M. Bayat Makou 44, L.T. Bayer 47, L.T. Bazzano Hurrell 31, T. Beau 90, J.Y. Beaucamp 90, S. Beauceron 128, P.H. Beauch emin 161, P. Bechtle 25, H.P. Beck 20, K. Becker 169, A.J. Beddall 80, V.A. Bednyakov 38, C.P. Bee 148, L.J. Beemster 16, M. Begalli 81d, M. Begel 30, J.K. Behr 47, J.F. Beirer 37, F. Beisiegel 25, I.B. Belean 28d, M. Belfkir 83c, G. Bella 154, L. Bellagamba 24b, A. Bellerive 35, C.D. Bellgraph 67, P. Bellos 21, I. Benaoumeur 21, D. Benchekroun 36a, F. Bendebba 36a, Y. Benhammou 154, K.C. Benkendorf er 167, L. Beresford 47, M. Beretta 52, E. Bergeaas Kuutmann 163, N. Berger 4, B. Bergmann 133, J. Beringer 18a, M. Berkat 136, G. Bernardi 5, C. Bernius 146, F.U. Bernlochner 25, A. Berrocal Guardia 13, T. Berry 95, P. Berta 134, A. Berti 131a, R. Bertrand 102, S. Bethke 110大香线蕉 A. Betti {sup}74a,74b, T.F. Beumker {sup}173, A.J. Bevan {sup}94, L. Bezio {sup}55, N.K. Bhalla {sup}53, M. Bhamjee {sup}34h, S. Bharthuar {sup}110, S. Bhatta {sup}148, S. Bhattacharya {sup}44, P. Bhattarai {sup}146, Z.M. Bhatti {sup}118, K.D. Bhide {sup}164, R.M. Bianchi {sup}130, O. Biebel {sup}109, M. Biglietti {sup}76a, P. Bijl {sup}53, C.S. Billingsley {sup}44, Y. Bimgdi {sup}36f, M. Bindi {sup}54, A. Bingham {sup}173, A. Bingul {sup}22b, C. Bini {sup}74a,74b, M. Biros {sup}134, S. Biryukov {sup}149, T. Bisanz {sup}48, E. Bisceglie {sup}24b,24a, J.P. Biswal {sup}135, D. Biswas {sup}144, M. Biyabi {sup}14, I. Bloch {sup}47, A. Blue {sup}58, U. Blumenschein {sup}94, V.S. Bobrovnikov {sup}38, L. Boccardo {sup}56b,56a, M. Boehler {sup}53, D. Bogavac {sup}13, L.S. Boggia {sup}128, V. Boisvert {sup}95, P. Bokan {sup}163, T. Bold {sup}85a, M. Bomben {sup}5, M. Bona {sup}94, M. Boonekamp {sup}136, A.G. Borbély {sup}58, G. Borissov {sup}91, A. Borkar {sup}168, D. Bortoletto {sup}127, M. Borysova {sup}171, D. Boscherini {sup}24b, M. Bosman {sup}13, K. Bouaouda {sup}36a, L. Boudet {sup}136, J. Boudreau {sup}130, E.V. Bouhova-Thacker {sup}91, D. Boumediene {sup}40, R. Bouquet {sup}56b,56a, A. Boveia {sup}120, D. Boye {sup}30, I.R. Boyko {sup}38, L. Bozianu {sup}37, J. Bracinik {sup}21, N. Brahimi {sup}4, G. Brandt {sup}173, O. Brandt {sup}33, B. Brau {sup}103, R. Brener {sup}171, L. Brenner {sup}116, R. Brenner {sup}163, M. Bressan {sup}41, S. Bressler {sup}171, M. Brettell {sup}96, G. Brianti {sup}116, D. Britton {sup}58, D. Britzger {sup}110, I. Brock {sup}25, R. Brock {sup}107, H. Bronson {sup}129, G. Brooijmans {sup}41, A.J. Brooks {sup}67, E.M. Brooks {sup}159b, E. Brost {sup}30, L.M. Brown {sup}167,159a, L.E. Bruce {sup}60, P.A. Bruckman de Renstrom {sup}86, B. Brüers {sup}47, A. Bruni {sup}24b, G. Bruni {sup}24b, D. Brunner {sup}46a,46b, M. Bruschi {sup}24b, N. Bruscino {sup}74a,74b, T. Buanes {sup}17, Q. Buat {sup}140, D. Buchin {sup}110, A.G. Buckley {sup}58, J. Bucko {sup}134, M. Buhring {sup}49, O. Bulekov {sup}80, B.A. Bullard {sup}146, T.O. Buratovich {sup}90, S. Burdin {sup}92, C.D. Burgard {sup}48, A.M. Burger {sup}89, B. Burghgrave {sup}8, J. Burleson {sup}164, J.C. Burzynski {sup}121, V. Büscher {sup}100, P.J. Bussey {sup}58, O. But {sup}25, J.M. Butler {sup}26, C.M. Buttar {sup}58, J.M. Butterworth {sup}96, P. Butti {sup}37, W. Buttinger {sup}135, C.J. Buxo Vazquez {sup}107, A.R. Buzykaev {sup}38, S. Cabrera Urbán {sup}165, L. Cadamuro {sup}65, H. Cai {sup}37, Q. Cai {sup}15, Y. Cai {sup}24b,112c,24a, Y. Cai {sup}112a, M.A. Cairo {sup}129, V.M.M. Cairo {sup}37, O. Cakir {sup}3a, N. Calace {sup}37, P. Calafiura {sup}18a, G. Calderini {sup}128, P. Calfayan {sup}35, L. Calic {sup}98, G. Callea {sup}58, L.P. Caloba {sup}81b, D. Calvet {sup}40, S. Calvet {sup}40, R. Camacho Toro {sup}128, S. Camarda {sup}37, D. Camarero Munoz {sup}27, P. Camarri {sup}75a,75b, C. Camincher {sup}37, M. Campanelli {sup}96, A. Camplani {sup}42, V. Canale {sup}71a,71b, A.C. Canbay {sup}3a, E. Canonero {sup}95, J. Cantero {sup}165, F. Capocasa {sup}27, P. Cappelli {sup}27, M. Capua {sup}43b,43a, A. Carbone {sup}70a,70b, R. Cardarelli {sup}75a, J.C.J. Cardenas {sup}8, M.P. Cardiff {sup}27, G. Carducci {sup}43b,43a, T. Carli {sup}37, G. Carlino {sup}71a, J. Carlotto {sup}13, B.T. Carlson {sup}130,q, E.M. Carlson {sup}167, L. Carminati {sup}70a,70b, A. Carnelli {sup}4, M. Carnesale {sup}37, S. Caron {sup}115, E.M. Carpenter {sup}106, E. Carquin {sup}138, I.B. Carr {sup}105, S. Carrà {sup}72a,72b, G. Carratta {sup}24b,24a, C. Carrion Martinez {sup}165, A.M. Carroll {sup}124, N. Cartalade {sup}40, M.P. Casado {sup}13,h, A. Casali {sup}58, P. Casolaro {sup}71a,71b, F. Cassinese {sup}90, F. Castiglioni {sup}73a,73b, W.R. Castiglioni {sup}39, F.L. Castillo {sup}4, V. Castillo Gimenez {sup}165, N.F. Castro {sup}131a,131e, A. Catinaccio {sup}37, J.R. Catmore {sup}126, T. Cavaliere {sup}4, V. Cavaliere {sup}30, E. Celebi {sup}80, S. Cella {sup}30, V. Cepaitis {sup}55, K. Cerny {sup}123, A.S. Cerqueira {sup}81a, A. Cerri {sup}73a,aq, L. Cerrito {sup}75a,75b, F. Cerutti {sup}18a, B. Cervato {sup}70a, A. Cervelli {sup}24b, G. Cesarini {sup}52, S.A. Cetin {sup}80, V.C. Chabalala {sup}34, P.M. Chaballat {sup}128, R. Chakhrappah {sup}65, S. Chakraborty {sup}169, A. Chambers {sup}60, J. Chan {sup}18a, J.D. Chapman {sup}33, E. Chapple {sup}136, D.G. Charlton {sup}21, C. Chauhan {sup}132, Y. Che {sup}112a, S. Chekanov {sup}6, G.A. Chelkov {sup}38,a, H. Chen {sup}30, J. Chen {sup}145, M. Chen {sup}59, S. Chen {sup}87, S.J. Chen {sup}112a, X. Chen {sup}141a, X. Chen {sup}15,aj, Z. Chen {sup}61, C.L. Cheng {sup}146, H.C. Cheng {sup}63a, S. Cheong {sup}146, A. Cheplakov {sup}38, E. Cherepanova {sup}116, E. Cheu {sup}7, K. Cheung {sup}64, L. Chevalier {sup}136, G. Chiarelli {sup}73a, G. Chiodini {sup}69a, A.S. Chisholm {sup}21, J.L. Chisholm {sup}166, A. Chitan {sup}28b, M. Chitishvili {sup}165, M.V. Chizhov {sup}38,r, K. Chmiel {sup}76a,76b, K. Choi {sup}11, Y. Chou {sup}140, H. Choudhary {sup}145, E.Y.S. Chow {sup}115, G. Christou {sup}51, K.L. Chu {sup}171, M.C. Chu {sup}63a, Z. Chubinidze {sup}52, J. Chudoba {sup}132, J.J. Chwastowski {sup}86, D. Cieri {sup}110, K.M. Ciesla {sup}85a, J.P. Cifuentes Salazar {sup}23b, V. Cindro {sup}93, A. Ciocio {sup}18a, F. Cirotto {sup}71a,71b, Z.H. Citron {sup}171,k, M. Citterio {sup}70a, D.A. Ciubotaru {sup}28b, A. Clark {sup}55, P.J. Clark {sup}51, N. Clarke Hall {sup}37, C. Clarry {sup}158, S.E. Clawson {sup}37, C. Clement {sup}46a,46b, L. Clissa {sup}24b,24a, Y. Coadou {sup}102 M. Cobal 68a,68c, A. Coccaro 56b, M.G. Cochran Branson 140, R.F. Coelho Barrue 131a, R. Coelho Lopes De Sa 103, S. Coelli 70a, M.M. Cohen 129, L.S. Colangeli 158, B. Cole 41, P. Collado Soto 99, J. Collot 59, M.R. Coluccia 69a, I. Combes 65, P. Conde Muiño 131a,131g, L.H.J. Condren 162, M.P. Connell 34c, S.H. Connell 34c, E.I. Conroy 161, M. Contreras Cossio 11, F. Conventi 71a,al, A.M. Cooper-Sarkar 127, L. Corazzina 74a,74b, F.A. Corchia 24b,24a, A. Cordeiro Oudot Choi 140, L.D. Corpe 40, M. Corradi 74a,74b, F. Corriveau 104,ab, A. Cortes-Gonzalez 156, M.J. Costa 165, F. Costanza 4, D. Costanzo 142, J. Couthures 4, G. Cowan 95, K. Cranmer 172, L. Cremer 48, D. Cremonini 24b,24a, S. Crépé-Renaudin 59, F. Crescioli 128, T. Cresta 72a,72b, M. Cristinziani 144, M. Cristoforetti 77a,77b, T.M. Critchley 55, E. Critelli 96, A.S. Cruz 67, M. Cucinotta 73a,73b, A. Cueto 99, H. Cui 96, Z. Cui 7, B.M. Cunnett 149, W.R. Cunningham 58, E. Cuppini 110, F. Curcio 14, J.R. Curran 51, J.V. Da Fonseca Pinto 81b, C. Da Via 101, W. Dabrowski 85a, T. Dado 37, S. Dahbi 151, T. Dai 106, D. Dal Santo 20, C. Dallapiccola 103, M. Dam 42, G. D’amen 30, V. D’Amico 109, J.R. Dandoy 35, M. D’Andrea 56b,56a, D. Dannheim 37, G. D’anniballe 73a,73b, M. Danninger 145, V. Dao 148, G. Darbo 56b, F. Dattola 47, S. D’Auria 70a,70b, A. D’Avanzo 71a,71b, T. Davidek 134, J. Davidson 169, I. Dawson 94, K. De 8, C. De Almeida Rossi 158, S. De Castro 24b,24a, N. De Groot 115, P. de Jong 116, H. De la Torre 117, A. De Maria 112a, S. De Miranda Rimes 81d, A. De Salvo 74a, U. De Sanctis 75a,75b, A. De Santo 149, J.B. De Vivie De Regie 59, K.G. De Vries 116, J. Debevc 93, D.V. Dedovich 38, J. Degens 92, A.M. Deiana 44, J. Del Peso 99, L. Delagrange 27, F. Deliot 136, C.M. Delitzsch 48, M. Della Pietra 71a,71b, S.R.J. Della Santa 101, D. Della Volpe 55, A. Dell’Acqua 37, L. Dell’Asta 70a,70b, M. Delmastro 4, C.C. Delogu 56b,56a, P.A. Delsart 59, S. Demers 174, M. Demichev 38, H. Denizli 22a,l, M.G. Depala 92, L. D’Eramo 40, D. Derendarz 86, L. Derin 56b,56a, F. Derue 128, P. Dervan 92,*, A.M. Desai 1, K. Desch 25, F.A. Di Bello 73a,73b, A. Di Ciaccio 75a,75b, L. Di Ciaccio 4, D. Di Croce 37, C. Di Donato 71a,71b, A. Di Girolamo 37, G. Di Gregorio 65, A. Di Luca 77a,77b, B. Di Micco 76a,76b, R. Di Nardo 76a,76b, K.F. Di Petrillo 39, M. Diamantopoulou 154, F.A. Dias 116, N. Dias Pereira Neto 81e, M.A. Diaz 138a,138b, A.R. Didenko 38, S.D. Diefenbacher 62a, E.B. Diehl 106, S. Díez Cornell 47, C. Diez Pardos 144, C. Dimitriadi 147, A. Dimitrievska 21, A. Dimri 148, Y. Ding 61, J. Dingfelder 25, T. Dingley 127, I-M. Dinu 28b, S.J. Dittmeier 62b, F. Dittus 37, M. Divisek 134, B. Dixit 92, F. Djama 102, T. Djobava 152b, C. Doglioni 101,98, A. Dohnalova 29a, Z. Dolezal 134, K.M. Dona 39, M. Donadelli 81d, B. Dong 107, J. Donini 40, A. D’Onofrio 71a,71b, M. D’Onofrio 92, J. Dopke 135, A. Doria 71a, N. Dos Santos Fernandes 131a, I.A. Dos Santos Luz 81e, P. Dougan 44, M.T. Dova 90, A.T. Doyle 58, M.P. Drescher 54, E. Drey er 171, I. Drivas-Koulouris 10, M. Drnevich 118, D. Du 61, T. Du 39, T.A. du Pree 116, Z. Duan 112a, M. Dubau 96, F. Dubinin 38, M. Dubovsky 29a, E. Duchovni 61, G. Dueck 116, P.K. Duckett 96, O.A. Ducu 28b, D. Duda 51, A. Dudarev 37, M.M. Dudek 86, E.R. Duden 27, M. D’uffizi 101, L. Duflot 65, M. Dührssen 37, J. Duminica 28g, A.E. Dumitriu 28b, M. Dunford 62a, T. Duong 4, A. Duperrin 102, A.F. Duque Bran 40, G. Duraikandan 134, H. Duran Yildiz 3a, A. Durg lishvili 152b, G.I. Dyckes 18a, M. Dyndal 85a, B.S. Dziedzic 37, G.H. Eberwein 127, B. Eckerova 29a, J.C. Egan 96, S. Eggebrecht 54, E. Egidio Purcino De Souza 81e, G. Eigen 17, K. Einsweiler 18a, T. Ekelof 163, P.A. Ekman 98, S. El Farkh 36b, Y. El Ghazali 61, H. El Jarrari 104, A. El Moussaouy 36a, I. Elbaz 154, D. Elitez 37, M. Ellert 163, F. Ellinghaus 95, T.A. Elliot 95, J. Elmsheuser 30, M. Elsawy 83b, M. Elsing 37, S. Emami 105, D. Emelyanov 135, Y. Enari 82, C. Engel 100, S. Epari 1, D. Ernani Martins Neto 86, F. Ernst 37, M. Escalier 65, C. Escobar 165, S. Esteva n De Paula 81c, E. Etzion 154, G. Evans 131a,131b, H. Evans 67, L.S. Evans 47, S. Ezz arqtouni 36a, F. Fabbri 24b,24a, L. Fabbri 24b,24a, G. Facini 96, V. Fadeyev 137, D. Fakoudis 100, S. Falciano 74a, L.F. Falda Ulhoa Coelho {sup}27, F. Fallavollita {sup}110, G. Falsetti {sup}43b,43a, J. Faltova {sup}134, K.Y. Fan {sup}63b, Y. Fan {sup}14, Y. Fang {sup}114,112c, M. Fanti {sup}70a,70b, M. Faraj {sup}68a,68c, Z. Farazpay {sup}97, A. Farbin {sup}8, A. Farilla {sup}76a, K. Farman {sup}151, J.N. Farr {sup}174, M.S. Farrington {sup}60, S.M. Farrington {sup}135,51, F. Fassi {sup}36e, D. Fassouliotis {sup}9, L. Fayard {sup}65, G. Fazzino {sup}62b, P. Federic {sup}134, P. Federicova {sup}132, M. Feickert {sup}172, L. Feligioni {sup}102, D.E. Fellers {sup}18a, C. Feng {sup}113b, Y. Feng {sup}14, Z. Feng {sup}65, B. Fernandez Barbadillo {sup}91, P. Fernandez Martinez {sup}66, C. Fernandez Ruiz {sup}33, J. Ferrando {sup}91, A. Ferrari {sup}163, P. Ferrari {sup}116,115, R. Ferrari {sup}72a, D. Ferrere {sup}55, C. Ferretti {sup}106, M.P. Fewell {sup}1, D. Fiacco {sup}55, F. Fiedler {sup}100, P. Fiedler {sup}133, K. Figueredo Rodriguez {sup}165, S. Filimonov {sup}38, M.S. Filip {sup}28b,s, M. Filipig {sup}68a,68c, A. Filipčič {sup}93, E.K. Filmer {sup}159a, F. Filthaut {sup}115, M.C.N. Fiolhais {sup}131a,131c,c, L. Fiorini {sup}165, F. Fischer {sup}100, W.C. Fisher {sup}107, T. Fitschen {sup}101, I. Fleck {sup}144, P. Fleischmann {sup}106, T. Flick {sup}173, M. Flores {sup}34d,ag, L.R. Flores Castillo {sup}63a, M. Foll {sup}126, F.M. Follega {sup}77a,77b, N. Fomin {sup}33, A. Formica {sup}136, M. Fornasiero {sup}149, A.C. Forti {sup}101, N. Forti {sup}24b,24a, E. Fortin {sup}102, A.W. Fortman {sup}18a, L. Foster {sup}18a, L. Fountas {sup}9, H. Fox {sup}91, P. Francavilla {sup}73a,73b, S. Francescato {sup}60, S. Franchellucci {sup}20, M. Franchini {sup}24b,24a, S. Franchino {sup}62a, D. Francis {sup}37, L. Franco {sup}47, L. Franconi {sup}47, M. Franklin {sup}60, G. Frattari {sup}37, Y.Y. Frid {sup}154, N. Fritzsche {sup}37, A. Froch {sup}55, D. Froidevaux {sup}37, J.A. Frost {sup}135, Y. Fu {sup}107, S. Fuenzalida Garrido {sup}138g, Y.C. Fujikake {sup}137, M. Fujimoto {sup}148, K.Y. Fung {sup}63a, E. Furtado De Simas Filho {sup}81e, M. Furukawa {sup}156, M. Fuste Costa {sup}47, P. Fuste Martin {sup}13, J. Fuster {sup}165, A. Gaa {sup}54, A. Gabrielli {sup}24b,24a, A. Gabrielli {sup}158, G. Gagliardi {sup}56b,56a, L.G. Gagnon {sup}146, S. Gaid {sup}68a,68b, S. Galantzan {sup}154, I. Gales Alves Correia Pinto {sup}144, J. Gallagher {sup}1, E.J. Gallas {sup}127, A.L. Gallen {sup}163, B.J. Gallop {sup}135, K.K. Gan {sup}120, Y. Gao {sup}51, Z. Gao {sup}112a, A. Garabaglu {sup}140, F.M. Garay Walls {sup}138a,138b, C. García {sup}165, A. Garcia Alonso {sup}116, A.G. Garcia Caffaro {sup}174, J.E. García Navarro {sup}165, M.A. Garcia Ruiz {sup}23b, M. Garcia-Sciveres {sup}18a, G.L. Gardner {sup}129, R.W. Gardner {sup}39, N. Garelli {sup}161, R.B. Garg {sup}146, J.M. Gargan {sup}33, C.A. Garner {sup}158, C.M. Garvey {sup}34a, V.K. Gassmann {sup}161, G. Gaudio {sup}72a, A.J. Gavin {sup}94, J. Gavranovic {sup}93, I.L. Gavrilenko {sup}131a, C. Gay {sup}166, G. Gaycken {sup}124, A. Gekow {sup}120, C. Gemme {sup}56b, M.H. Genest {sup}59, A.D. Gentry {sup}114, S. George {sup}95, T. Geralis {sup}45, A.A. Gerwin {sup}121, P. Gessinger-Befurt {sup}37, M. Ghani {sup}169, K. Ghorbanian {sup}94, A. Ghosal {sup}144, A. Ghosh {sup}162, B. Giacobbe {sup}24b, S. Giagu {sup}74a,74b, A. Giannini {sup}61, S.M. Gibson {sup}95, D.T. Gil {sup}85b, B.J. Gilbert {sup}41, D. Gillberg {sup}35, D.M. Gingrich {sup}2,ak, M.P. Giordani {sup}68a,68c, P.F. Giraud {sup}136, G. Giugliarelli {sup}68a,68c, D. Giugni {sup}70a, F. Giuli {sup}75a,75b,am, I. Gkialas {sup}9,i, B.C. Gladwyn {sup}127, C. Glasman {sup}99, M. Glazewska {sup}20, R.M. Gleason {sup}162, G. Glemžа {sup}47, I. Gnesi {sup}24b,24a,an, Y. Go {sup}30, M. Goblirsch-Kolb {sup}37, B. Gocke {sup}48, D. Godin {sup}108, B. Gokturk {sup}22a, S. Goldfarb {sup}105, T. Golling {sup}55, M.G.D. Gololo {sup}34c, A. Golub {sup}140, J.P. Gombas {sup}107, A. Gomes {sup}131a,131b, G. Gomes Da Silva {sup}144, A.J. Gomez Delegido {sup}37, R. Gonçalo {sup}51, A. Gongadze {sup}152c, F. Gonnella {sup}21, J.L. Gonski {sup}146, R.Y. González Andana {sup}51, S. González de la Hoz {sup}165, M.V. Gonzalez Rodrigues {sup}47, R. Gonzalez Suarez {sup}163, S. Gonzalez-Sevilla {sup}55, L. Goossens {sup}37, B. Gorini {sup}37, E. Gorini {sup}69a,69b, A. Gorišek {sup}93, T.C. Gosart {sup}129, A.T. Goshaw {sup}50, M.I. Gostkin {sup}38, S. Goswami {sup}122, C.A. Gottardo {sup}37, S.A. Gotz {sup}109, M. Gouighri {sup}36b, A.G. Goussious {sup}140, N. Govender {sup}34c, R.P. Grabarczyk {sup}127, I. Grabowska-Bold {sup}85a, K. Graham {sup}35, E. Gramstad {sup}126, S. Grancagnolo {sup}69a,69b, P.M. Gravila {sup}28f, F.G. Gravili {sup}69a,69b, H.M. Gray {sup}18a, M. Greco {sup}110, M.J. Green {sup}1, C. Grefe {sup}25, A. Grefsruid {sup}37, I.M. Gregor {sup}47, K.T. Greif {sup}162, P. Grenier {sup}146, S.G. Grewe {sup}110, K. Grimm {sup}32, S. Grinstein {sup}13,x, E. Gross {sup}171, J. Grosse-Knetter {sup}54, L.H. Grossman {sup}18b, L. Guan {sup}106, G. Guerrieri {sup}37, R. Guevara {sup}126, R. Gugel {sup}100, A. Guida {sup}19, S. Guindon {sup}37, F. Guo {sup}114,112c, J. Guo {sup}141a, L. Guo {sup}47, L. Guo {sup}112b, Y. Guo {sup}106, Y. Guo {sup}41, A. Gupta {sup}48, R. Gupta {sup}130, S. Gupta {sup}27, S. Gurbuz {sup}25, S.S. Gurdasani {sup}47, G. Gustavino {sup}74a,74b, P. Gutierrez {sup}121, L.F. Gutierrez Zagazeta {sup}129, M. Gutsche {sup}49, C. Gutschow {sup}96, W. Guérin {sup}89` C. Gwenlan 127, C.B. Gwilliam 92, E.S. Haaland 126, A. Haas 118, M. Habedank 58, C. Haber 18a, R.J. Haberle 171, H.K. Hadavand 8, A. Haddad 40, A. Hadef 49, K.E. Haeussler 53, A.I. Hagan 91, J.J. Hahn 144, M. Haleem 168, J. Haley 122, G.D. Hallewell 102, J.A. Hallford 47, H. Hamdaoui 163, M. Hamer 144, S.E.D. Hammoud 65, E.J. Hampshire 95, L. Han 112a, L. Han 61, S. Han 14, K. Hanagaki 82, M. Hance 137, D.A. Hangal 41, H. Hanif 145, M.D. Hank 129, J.B. Hansen 42, P.H. Hansen 42, T. Harenberg 173, S. Harkusha 175, M.L. Harris 103, Y.T. Harris 25, J. Harrison 13, P.F. Harrison169, M.L.E. Hart 96, N.M. Hartman 110, N.M. Hartmann 109, R.Z. Hasan 95,135, Y. Hasegawa 143, D. Hashimoto 111, F. Haslbeck 37, S. Hassan 126, R. Hauser 107, M. Haviernik 134, C.M. Hawkes 21, R.J. Hawkings 37, Y. Hayashi 156, D. Hayden 107, R.L. Hayes 116, C.P. Hays 127, J.M. Hays 94, H.S. Hayward 92, M. He 14,112c, Y. He 47, Y. He 96, V. Hedberg 98, J. Heilman 35, S. Heim 47, T. Heim 18a, J.J. Heinrich 124, L. Heinrich 110, J. Hejbal 132, M. Helbig 49, A. Held 172, S. Hellesund 17, C.M. Helling 166, F.N.E. Henry 58, H. Herde 98, Y. Hernández Jiménez 148, M. Hernandez Sanz 116, G. Herten 53, R. Hertenberger 109, L. Hervas 37, T.C. Herwig 106, M.E. Hesping 100, N.P. Hessey 159a, J. Hessler 110, R. Hicks 129, M. Hidaoui 113b, N. Hidic 134, E. Hill 158, T.S. Hillersoy 17, S.J. Hillier 21, J.R. Hinds 107, F. Hinterkeuser 25, M. Hirose 125, S. Hirose 170, D. Hirschbuehl 173, B. Hiti 93, J. Hobbs 148, R. Hobincu 28e, N. Hod 171, A.M. Hodges 164, M.C. Hodgkinson 142, B.H. Hodkinson 37, A. Hoecker 37, D.D. Hofer 106, J. Hofer 165, J. Hofner 100, M. Holzbock 110, L.B.A.H. Hommels 33, V. Homsak 127, J.J. Hong 67, T.M. Hong 130, R. Honscheid 127, B.H. Hooberman 164, W.H. Hopkins 6, M.C. Hoppesch 164, Y. Horii 111, M.E. Horstmann 110, M.M. Horzela 54, S. Hou 151, M.R. Housenga 164, J. Howarth 58, J. Hoya 6, M. Hrabovsky 123, T. Hryn’ova 4, P.J. Hsu 64, S.-C. Hsu 140, T. Hsu 65, M. Hu 18a, P. Hu 63b, Q. Hu 61, S. Huang 33, X. Huang 14,112c, Y. Huang 134, Y. Huang 112b, Y. Huang 14, Z. Huang 65, Z. Hubacek 133, F. Huegging 25, T.B. Huffman 127, M. Hufnagel Maranha De Faria 81a, C.A. Hugli 47, M. Huhtinen 37, S.K. Huiberts 17, R. Hulsken 104, C.E. Hultquist 18a, D.L. Humphreys 103, N. Huseynov 12, J. Huston 107, B. Huth 37, J. Huth 60, L. Huth 47, R. Hyneman 7, G. Iacobucci 55, G. Iakovidis 30, L. Iconomidou-Fayard 65, J.P. Iddon 165, P. Iengo 71a,71b, Y. Iiyama 156, T. Iizawa 156, Y. Ikegami 82, D. Iliadis 155, N. Ilic 158, H. Imam 36a, G. Inacio Goncalves 81d, S.A. Infante Cabanas 138c, J.M. Inglis 94, G. Introzzi 72a,72b, M. Iodice 76a, V. Ippolito 74a,74b, R.K. Irwin 92, M. Ishino 156, W. Islam 172, C. Issever 19, O.N. Istaıtia 65,b, S. Istin 22a,as, K. Itabashi 125, H. Ito 170, R. Iuppa 77a,77b, A. Ivina 171, F. Ivone 37, S. Izumiyama 111, V. Izzo 71a, P. Jacka 133, P. Jackson 1, P.R. Jacobson 50, P. Jain 47, K. Jakobs 53, J. Jamieson 58, W. Jang 156, S. Jankovych 53, B.K. Jashal 135, M. Javurkova 103, P. Jawahar 101, L. Jeanty 124, J. Jejelava 152a,ae, P. Jenni 53,f, L. Jerala 93, C.E. Jessiman 35, H. Jia 166, J. Jia 148, K. Jia 146, X. Jia 110,112c, P. Jiang 51, D.G. Jiang 164, Q. Jiang 63, S. Jiggins 47, M. Jimenez 165, J. Jimenez Pena 13, S. Jin 112a, A. Jinaru 28b, O. Jinnouchi 139c, P. Johansson 142, K.A. Johns 7, J.W. Johnson 137, F.A. Jolly 47, D.M. Jones 149, E. Jones 47, R.W.L. Jones 91, T.J. Jones 92, H.L. Joos 37, R. Joshi 120, J. Jovcicevic 16, X. Ju 18a, J.J. Junggeburth 37, T. Junkermann 62a, A. Juste Rozas 13,x, M.K. Juzek 86, S. Kabana 138f, A. Kaczmarska 86, S.A. Kadir 146, M. Kado 110, P. Kafle 107, H. Kagan 120, M. Kagan 146, A. Kahn 129, C. Kahra 100, T. Kaji 156, E. Kajomovitz 153, N. Kakati 171, N. Kakoty 13, S. Kandel 8, E. Kanellaki 45, N. Kanellos 9, S. Kang 50, D. Kar 34,j, E. Karentzos 25, K. Karki 8, O. Karkout 16, S.N. Karpov 3, Z.M. Karpova 9, V. Kartvelishvili 91,152b, E. Kasimi 155, K. Katsarov 47, J. Katzy 47, S. Kaur 35, R. Kavak 37, K. Kavatsyuk 143, M.P. Kawale 121, C. Kawamoto 87, E.F. Kay 37, S. Kazakos 107, K. Kazakova 102, J.M. Keaveney 34a, R. Keeler 167, G.V. Kehris 60, J.S. Keller 35, J.M. Kelly 167, J.I. Kelsey 164 K. Kemp{sup}1, J.J. Kempster{sup}149, O. Kepka{sup}132, J. Kerr{sup}159b, B.P. Kerridge{sup}135, B.P. Kerševan{sup}93, L. Keszeghova{sup}29a, M.B. Khan{sup}93, R.A. Khan{sup}130, A. Khanov{sup}122, M. Kholodenko{sup}131a, T.J. Khoo{sup}19, G. Khoriauli{sup}168, Y. Khoulaki{sup}36a, Y.A.R. Khwaira{sup}128, D. Kim{sup}6, D.W. Kim{sup}18b, Y.K. Kim{sup}39, N. Kimura{sup}96, M.K. Kingston{sup}54, F. Kirfel{sup}25, J. Kirk{sup}135, A.E. Kiryunin{sup}110,

S. Kita{sup}156, O. Kivernyk{sup}25, J. Klas{sup}25, M. Klassen{sup}37, C. Klein{sup}35, M.H. Klein{sup}44, U. Klein{sup}92, A. Klimentov{sup}30, P. Kluit{sup}116, S. Kluth{sup}110, E. Kneringer{sup}78, T.M. Knight{sup}158, A. Knue{sup}48,

M. Kobel{sup}49, D. Kobylianskii{sup}171, S.F. Koch{sup}37, M. Kocian{sup}146, P. Kodyš{sup}134, D.M. Koeck{sup}124, T. Koffas{sup}35, K. Kojima{sup}82, O. Kolay{sup}49, I. Koletsou{sup}4, T. Komarek{sup}86, S. Kondo{sup}156,

K. Köneke{sup}54, A.X.Y. Kong{sup}1, T. Kono{sup}119, N. Konstantinidis{sup}96, P. Kontaxakis{sup}55, B. Konya{sup}98, R. Kopeliansky{sup}41, S. Koperny{sup}85a, R. Koppenhofer{sup}53, I. Kopsalis{sup}10, K. Korcyl{sup}86,

K. Kordas{sup}155,d, A. Korn{sup}96, S. Korn{sup}54, I. Korolkov{sup}13, O. Kortner{sup}110, S. Kortner{sup}110, W.H. Kostecka{sup}117, M. Kostov{sup}29a, V.V. Kostyukhin{sup}144, A. Kotsokechagia{sup}37, A. Kotwal{sup}50,

A. Koulouris{sup}37, A. Kourkoumeli-Charalampidi{sup}72a,72b, O. Kovanda{sup}124, R. Kowalewski{sup}167, W. Kozanecki{sup}124, G. Kramberger{sup}93, P. Kramer{sup}25, A. Krasznahorkay{sup}103, A.C. Kraus{sup}117,

J.W. Kraus{sup}173, J.A. Kremer{sup}85a, N.B. Krengel{sup}144, T. Kresse{sup}158, L. Kretschmann{sup}173, J. Kretzschmar{sup}92, P. Krieger{sup}158, K. Krizka{sup}21, K. Kroeninger{sup}48, H. Kroha{sup}110, J. Kroll{sup}132,

J. Kroll{sup}129, K.S. Krowpman{sup}107, U. Kruchonak{sup}38, H. Krüger{sup}25, N. Krumnack{sup}79, J. Krupa{sup}146, M.C. Kruse{sup}50, O. Kuchinskaia{sup}38, S. Kuday{sup}3a, S. Kuehn{sup}37, R. Kuesters{sup}53, R. Kugo{sup}125,

T. Kuhl{sup}47, Y. Kulchitsky{sup}38, S. Kuleshov{sup}138d,138b, J. Kull{sup}1, E.V. Kumar{sup}109, M. Kumar{sup}34j, N. Kumari{sup}47, P. Kumari{sup}159b, T. Kumita{sup}157, A. Kupco{sup}132, O. Kuprash{sup}53, H. Kurashige{sup}84,

L.L. Kurchaninov{sup}159a, O. Kurdysh{sup}4, M. Kuze{sup}139, A.K. Kvam{sup}103, J. Kvita{sup}123, A. Kyprianou{sup}51, N.G. Kyriacou{sup}140, M. Laassiri{sup}30, C. Lacasta{sup}165, H. Lacker{sup}19, D. Lacour{sup}128,

E. Ladygin{sup}38, A. Lafarge{sup}40, B. Laforge{sup}128, T. Lagouri{sup}174, F.Z. Lahbabi{sup}36a, S. Lai{sup}54, W.S. Lai{sup}96, I.K. Lakomiec{sup}54, S. Lammers{sup}67, W. Lampl{sup}7, C. Lampoudis{sup}155,

G. Lamprinoudis{sup}168, A.N. Lancaster{sup}117, U. Landgraf{sup}53, M.P.J. Landon{sup}94, V.S. Lang{sup}53, A.J. Lankford{sup}162, F. Lanni{sup}37, C.S. Lantz{sup}164, K. Lantzsch{sup}25, A. Lanza{sup}72a,

M. Lanzac Berrocal{sup}165, T. Lari{sup}70a, D. Larsen{sup}17, L. Larson{sup}11, F. Lasagni Manghi{sup}24b, M. Lassnig{sup}37, H.C. Lau{sup}167, S.D. Lawlor{sup}142, R. Lazaridou{sup}162, M. Lazzaroni{sup}70a,70b,

E.T.T. Le{sup}162, H.D.M. Le{sup}107, E.M. Le Boulicaut{sup}174, D.O. Le Guennec{sup}136, L.T. Le Pottier{sup}18a, B. Leban{sup}24b,24a, F. Ledroit-Guillon{sup}59, T.F. Lee{sup}159b, L.L. Leeuw{sup}34h, M. Lefebvre{sup}167,

C. Leggett{sup}18a, L.M. Lehmann{sup}116, W.A. Leight{sup}103, W. Leinonen{sup}115, A. Leisos{sup}155,t, M.A.L. Leite{sup}81c, C.E. Leitgeb{sup}19, R. Leitner{sup}134, E. Lelak{sup}134, K.J.C. Leney{sup}44, T. Lenz{sup}25,

S. Leone{sup}73a, C. Leonidopoulos{sup}51, A. Leopold{sup}147, J. LePage-Bourbonnais{sup}35, R. Lees{sup}65, C.G. Lester{sup}33, J.K. Leszczynska{sup}86, J. Levêque{sup}4, L.J. Levinson{sup}171, G. Levrini{sup}24b,24a,

M.P. Lewicki{sup}86, C. Lewis{sup}140, D.J. Lewis{sup}15, L. Lewitt{sup}142, A. Li{sup}9, B. Li{sup}106, C. Li{sup}110, H. Li{sup}113b, H. Li{sup}101, H. Li{sup}15, H. Li{sup}61, H. Li{sup}113b, J. Li{sup}106, J. Li{sup}14,

R. Li{sup}174, S. Li{sup}141b,141a, Y. Li{sup}14, Z. Li{sup}14,112c, Z. Li{sup}61, S. Liang{sup}141,112c, Z. Liang{sup}14, B. Liberti{sup}75a, G.B. Libotte{sup}81d, K. Lie{sup}63c, J. Lieber Marin{sup}81e, H. Lien{sup}67, H. Lin{sup}106,

S.F. Lin{sup}148, L. Linden{sup}109, R.E. Lindley{sup}7, J.H. Lindon{sup}37, J. Ling{sup}60, M. Linkert{sup}100, E. Lipeles{sup}129, A. Lipniacka{sup}17, A. Lister{sup}166, J.D. Little{sup}67, B. Liu{sup}113a, B.X. Liu{sup}112b,

D. Liu{sup}153, D. Liu{sup}137, E.H.L. Liu{sup}21, H. Liu{sup}112b, J.K.K. Liu{sup}118, K. Liu{sup}141b, K. Liu{sup}141b, M. Liu{sup}61, M.Y. Liu{sup}61, P. Liu{sup}113b, Q. Liu{sup}146, S. Liu{sup}148, X. Liu{sup}113b, Y. Liu{sup}112b,112c,

Y. Liu{sup}164, Y.L. Liu{sup}113b, Y.W. Liu{sup}61, Z. Liu{sup}65,j, S.L. Lloyd{sup}94, E.M. Lobodzinska{sup}47, P. Loch{sup}107, E. Lodhi{sup}158, K. Lohwasser{sup}15, E. Loiacono{sup}67, J.D. Lomas{sup}21, I. Longarini{sup}162,

R. Longo{sup}24b,24a,an, A. Lopez Solis{sup}113, N.A. Lopez-canelas{sup}7, N. Lorenzon{sup}4, A.M. Lory{sup}109, M. Losada{sup}83b, G. Löschcke Centeno{sup}4, X. Lou{sup}14,112c, P.A. Love{sup}91, H. Lu{sup}14,

M. Lu{sup}65, S. Lu{sup}129, Y.J. Lu{sup}151, H.J. Lubatti{sup}140,*, C. Luci{sup}74a,74b, F.L. Lucio Alves{sup}112a, ... J.A. Lue124, F. Luehring67, D. Lumb135, B.S. Lunday129, O. Lundberg147, J. Lunde37, N.A. Luongo6, M.S. Lutz158, A.B. Lux26, D. Lynn30, R. Lysak132, V. Lysenko133, E. Lytken98, V. Lyubushkin38, T. Lyubushkina38, M.M. Lyukova148, H. Ma30, K. Ma61, L.L. Ma113b, W. Ma61, Y. Ma113b, P.C. Machado De Abreu Farias81e, D. Macina37, R. Madar40, T. Madula96, J. Maeda84, S. Maeland17, T. Maeno30, P.T. Mafa34f, G. Magni65, H. Maguire142, M. Maheshwari33, V. Maiboroda65, G. Maineri70a,70b, A. Maio131a,131b,131d, A. Maiza168, K. Maj85a, O. Majersky47, S. Majewski124, A. Makita156, N. Makovec65, V. Maksimovic16, B. Malaescu128, J. Malamant126, Pa. Malecki86, F. Malek59,n, M. Mali93, D. Malito95, A. Maloizel5, A. Malvezzi Lopes81d, S. Malyukov38, J. Mamuzic93, G. Mancini52, M.N. Mancini27, G. Manco72a,72b, S.S. Mandarry149, I. Mandić93, L. Manhaes de Andrade Filho81a, I.M. Maniatis171, J. Manjarres Ramos89, D.C. Mankad171, A. Mann109, T. Manoussos100, M.N. Mantinan39, S. Manzoni37, L. Mao141a, X. Mapekula34c, A. Marantis155, R.R. Marcelo Gregorio1, G. Marchiori5, C. Marcon70a, J.P. Mariano20, E. Maricic16, M. Marinescu47, S. Marium47, M. Marjanovic121, A. Markhoos53, M. Markovitch65, M.K. Maroun103, M.C. Marr145, T.L. Marsault136, G.T. Marsden101, Z. Marshall18a, S. Marti-Garcia165, J. Martin96, T.A. Martin135, V.J. Martin51, B. Martin dit Latour17, F. Martina69a,69b, L. Martinelli74a,74b, V.I. Martinez Outschoorn103, P. Martinez Suarez37, S. Martin-Haugh135, G. Martinovicova134, V.S. Martoiu28b, A. Martone89, A.C. Martyniuk96, A. Marzin37, D. Mascione77a,77b, L. Masetti100, J. Masik101, A.L. Maslennikov38, S.L. Mason41, P. Massarotti71a,71b, P. Mastrandrea73a,73b, A. Mastroberardino43b,43a, R. Mastrofrancesco72a,72b, T. Masubuchi125, T.T. Mathew121, J. Matousek134, D.M. Mattern48, K. Mauer47, J. Maurer28b, T. Maurin58, B. Maček93, C. Mavungu Tsava102, A.E. May101, E. Mayer40, R. Mazini34j, S.M. Mazza137, E. Mazzeo37, J.P. Mc Gowan167, S.P. Mc Kee106, C.C. McCracken166, E.F. McDonald105, L.F. Mcelhinney91, J.A. Mcfayden149, R.P. McGovern167, R.P. Mckenzie8, D.J. Mclaughlin96, S.J. McMahon135, C.M. Mcpartland92, R.A. McPherson167,ab, S. Mehlhase109, A. Mehta92, K. Mekhemar83a, D. Melini165, B.R. Mellado Garcia14,ai, A.H. Melo54, F. Meloni47, A.M. Mendes Jacques Da Costa101, L. Meng91, S. Menke110, M. Mentink37, E. Meoni43b,43a, G. Mercado117, S. Merianos155, C. Merlassino68a,68c, C. Meroni70a,70b, J. Metcalfe6, A.S. Mete6, E. Meuser100, C. Meyer67, J-P. Meyer136, O. Mezhenska29b, Y. Miao112a, R.P. Middleton135, M. Mihovilovic65, L. Mijović51, G. Mikenberg171, M. Mikestikova132, V.M. Mikuniah, M. Mikuž93, H. Mildner100, A. Milic37, D.W. Miller39, E.H. Miller146, L.S. Miller159a, A. Milov171, D.A. Milstead46a,46b, T. Min112a, I.A. Minashvili152b, A.I. Mincer118, B. Mindur85a, M. Mineev38, L.M. Mir13, M. Miralles Lopez58, M. Mironova18a, M. Missio40, A. Mitra169, P. Mitra13, V.A. Mitsou165, P.S. Miyagawa94, R. Mizuhiki84, M. Mkr t chyan37, M. Mlinarevic96, T. Mlinarevic96, M. Mlynarikova134, L. Mlynarska85a, C. Mo141a, H. Möbius47, S. Mobius20, M.H. Mohamed Farouk114, S. Mohapatra41, M.F. Mohd Soberi51, S. Mohiuddin122, R. Mole21, L. Moleri171, U. Molinatti127, M.E. Mollerach31, L.G. Mollier20, L. Monaco37,58, B. Mondal132, S. Mondal134, K. Mönig47, E. Monnier102, L. Monsonis Romero165, A. Montella46a,46b, M. Montella120, F. Montereali76a,76b, F. Monticelli90, S. Monzani68a,68c, M.E.E. Moors25, A. Mor an cho Tarda42, N. Morange65, M. Moreno Llácer165, C. Moreno Martinez55, P. Morettini56b, S. Morgenstern62a, M. Morii62, M. Morinaga115, F. Morodei74a,74b, P. Moschovakos37, B. Moser53, M. Mosz ide152b, T. Mosk ale ts44, P. Mos k vitina115, C.J. Mosomane34b, J. Moss32, T. Motta Quoirin81d, A. Moussa36d, M. Moyal171,k, H. Moyano Gomez13, E.J.W. Moyse103, T.G. Mroz86, S. Muanza102, M. Mucha25, P. Mucha96, J. Mueller130, B.J. Mughal17, D. Muller144, G.A. Muller163, S.B. Mulligan55, A.J. Mullin{sup}33, J.J. Mullin{sup}50, A.C. Mullins{sup}44, A.E. Mulski{sup}60, D.P. Mungo{sup}158, D. Munoz Perez{sup}122, F.J. Munoz Sanchez{sup}101, J.M. Murnauer{sup}110, W.J. Murray{sup}169,135, E. Musajan{sup}61, M. Muškinja{sup}93, C. Mwewa{sup}47, A.J. Myers{sup}8, G. Myers{sup}106, M. Myska{sup}133, B.P. Nachman{sup}146, I.A. Nadas{sup}28d, K. Nagai{sup}127, K. Nagano{sup}82, R. Nagasaka{sup}156, J.L. Nagle{sup}30,ap, E. Nagy{sup}102, A.M. Nairz{sup}37, T. Nakagawa{sup}87, Y. Nakahama{sup}82, K. Nakamura{sup}82, A. Nandi{sup}62b, H. Nanjo{sup}125, E.A. Narayanan{sup}44, V.A. Narendran{sup}127, Y. Narukawa{sup}156, L. Nasella{sup}27, S. Nasri{sup}83c, C. Nass{sup}25, G. Navarro{sup}23a, A. Nayaz{sup}19, S. Nechaeva{sup}24b,24a, F. Nechansky{sup}132, A. Negri{sup}72a,72b, M. Negrini{sup}24b, C. Nellist{sup}116, C. Nelson{sup}104, K. Nelson{sup}106, S. Nemecek{sup}132, M. Nessi{sup}37,g, M.S. Neubauer{sup}164, J. Newell{sup}92, P.R. Newman{sup}21, Y.W.Y. Ng{sup}164, B. Ngair{sup}83b, H.D.N. Nguyen{sup}108, J.D. Nichols{sup}121, R. Nicolaidou{sup}136, J. Nielsen{sup}137, M. Niemeyer{sup}54, J. Niermann{sup}37, N. Nikiforou{sup}37, I. Nikolic-Audit{sup}128, P. Nilsson{sup}30, G. Ninio{sup}154, A. Nisati{sup}74a, D. Nishimura{sup}156, R. Nisius{sup}110, N. Nitika{sup}171, E.K. Nkadimeng{sup}34j, T. Nobe{sup}156, D. Noll{sup}146, T. Nommensen{sup}150, M.B. Norfolk{sup}142, B.J. Norman{sup}35, L.C. Nosler{sup}18a, M. Noury{sup}36a, J. Novak{sup}93, T. Novak{sup}93, P. Novotny{sup}171, R. Novotny{sup}133, L. Nozka{sup}123, K. Ntekas{sup}37, D. Ntounis{sup}146, N.M.J. Nunes De Moura Junior{sup}81b, J. Ocariz{sup}128, I. Ochoa{sup}131a, A. Odella Rodriguez{sup}13, S. Oerdek{sup}47, K. Ograbek{sup}85a, A. Ogrodnik{sup}86, A. Oh{sup}101, C.C. Ohm{sup}147, H. Oide{sup}82, M.L. Ojeda{sup}37, Y. Okumura{sup}156, L.F. Oleiro Seabra{sup}131a, I. Oleksiyuk{sup}55, G. Oliveira Correa{sup}13, D. Oliveira Damazio{sup}30, J.L. Oliver{sup}1, R. Omar{sup}67, A.P. O’Neill{sup}20, Y. Onoda{sup}139, A. Onofre{sup}131a,131e,e, P.U.E. Onyisi{sup}11, M.J. Oreglia{sup}39, D. Orestano{sup}76a,76b, R. Orlandini{sup}76a,76b, R.S. Orr{sup}158, L.M. Osojnak{sup}41, Y. Osumi{sup}111, G. Otero y Garzón{sup}31, H. Otono{sup}88, M. Ouchrif{sup}36d, F. Ould-Saada{sup}126, T. Ovsiannikova{sup}140, M. Owen{sup}58, R.E. Owen{sup}135, S.A. Oyeniran{sup}114, V.E. Ozcan{sup}22a, F. Ozturk{sup}86, N. Ozturk{sup}8, S. Ozturk{sup}80, H.A. Pacey{sup}127, K. Pachal{sup}159a, A. Pacheco Pages{sup}13, N. Pacifico{sup}37, C. Padilla Aranda{sup}13, G. Padovano{sup}74a,74b, S. Pagan Griso{sup}18a, L. Pagani{sup}75a,75b, J. Pampel{sup}25, D.K. Panchal{sup}11, J. Panda{sup}169, C.E. Pandini{sup}59, J.G. Panduro Vazquez{sup}135, H.D. Pandya{sup}1, H. Pang{sup}136, P. Pani{sup}47, G. Panizzo{sup}68a,68c, L. Panwar{sup}128,w, L. Paolozzi{sup}21, S. Parajuli{sup}164, A. Paramonov{sup}6, C. Paraskevopoulos{sup}52, S.R. Paredes Saenz{sup}51, A. Pareti{sup}72a,72b, K.R. Park{sup}41, T.H. Park{sup}110, F. Parodi{sup}56b,56a, J.A. Parsons{sup}41, A. Partridge{sup}137, U. Parzefall{sup}53, B.A. Paschen{sup}18a, B. Pascual Dias{sup}40, L. Pascual Dominguez{sup}99, E. Pasqualucci{sup}74a, S. Passaggio{sup}56b, F. Pastore{sup}95, P. Patel{sup}86, U.M. Patel{sup}50, J.R. Pater{sup}101, T. Pauly{sup}37, A. Paunovic{sup}16, F. Pauwels{sup}134, P. Pavzderin{sup}171, C.I. Pazos{sup}161, M. Pedersen{sup}126, R. Pedro{sup}131a, O. Penc{sup}132, C.C. Penelaud{sup}128, S. Peng{sup}15, G.D. Penn{sup}174, B.S. Peralva{sup}81d, A.P. Pereira Peixoto{sup}140, L. Pereira Sanchez{sup}146, D.V. Perepelitsa{sup}30,ap, G. Perera{sup}103, E. Perez Codina{sup}37, M. Perganti{sup}10, H. Pernegger{sup}37, S. Perrella{sup}74a,74b, K. Peters{sup}47, R.F.Y. Peters{sup}101, B.A. Petersen{sup}37, T.C. Petersen{sup}42, E. Petit{sup}102, V. Petousis{sup}133, A.R. Petri{sup}70a,70b, V.A. Petrovic{sup}96, T. Petru{sup}134, M. Pettee{sup}18a, A. Petukhov{sup}80, K. Petukhova{sup}37, R. Pezoa{sup}138g, L. Pezzotti{sup}24b,24a, G. Pezzullo{sup}174, L. Pfaffelmoser{sup}37, A.J. Pflieger{sup}78, T.M. Pham{sup}172, T. Pham{sup}105, P.W. Phillips{sup}135, G. Piacquadio{sup}148, E. Pianori{sup}18a, F. Piazza{sup}124, R. Pieg aia{sup}31, D. Pietreanu{sup}28b, A.D. Pilkington{sup}101, T. Pilusa{sup}34j, M. Pinamonti{sup}68a,68c, J.L. Pinfold{sup}2, G. Pinheiro Matos{sup}41, B.C. Pinheiro Pereira{sup}131a, J. Pinol Bel{sup}13, A.E. Pinto Pinoargote{sup}128, L. Pintucci{sup}68a,68c, K.M. Piper{sup}149, A. Pirttikoski{sup}55, D.A. Pizzi{sup}35, L. Pizzimento{sup}63b, A. Pizzini{sup}126, A. Plebani{sup}33, M.-A. Pleier{sup}30, V. Pleskot{sup}134, E. Plotnikova{sup}38, G. Poddar{sup}94, R. Poettgen{sup}98, N.A. Pohl{sup}120, S. Polacek{sup}134, G. Polesello{sup}72a, A. Poley{sup}145, A. Polini{sup}24b, C.S. Pollard{sup}169, Z.B. Pollock{sup}120, E. Pompa Pacchi{sup}121, N.I. Pond{sup}96, D. Pon omarenko{sup}67, L. Pontecorvo{sup}37, S. Popa{sup}28a, G.A. Popeneciu{sup}28a, A. Poreba{sup}37, D.M. Portillo Quintero{sup}159a, S. Possi{sup}133, M.A. Postill{sup}142, P. Postolache{sup}28c, K. Potamianos{sup}169, P.A. Potepa{sup}85a, I.N. Potrap{sup}38, C.J. Potter{sup}33, H. Potti{sup}150, J. Poveda{sup}165, M.E. Pozo Astigarraga{sup}37, R. Pozzi{sup}37, A. Prades Ibanez{sup}75a,75b, S.R. Pradhan{sup}142, J. Preston{sup}95 J. Pretel167, D. Price101, M. Primavera69a, L. Primomo68a,68c, M.A. Principe Martin99, R. Privara123, T. Procter85b, M.L. Proffitt140, N. Proklova129, K. Prokofiev63c, G. Proto110, J. Proudfoot6, M. Przybycien85a, W.W. Przygoda85b, A. Psallidas45, D. Pudzha52, P. Puhl57, H.I. Purnell1, D. Pyatiizbyantseva115, I.A. Qattan83a, J. Qian106, R. Qian107, D. Qichen127, Y. Qin13, T. Qiu51, A. Quadt54, M. Queitsch-Maitland101, G. Quetant55, R.P. Quinn166

D. Rafanoharana110, J.L. Rainbolt39, S. Rajagopalan30, E. Ramakoti38, L. Rambelli56b,56a, J. Ramirez Alfaro165, I.A. Ramirez-Berend35, K. Ran106,112c, S.D. Randles92, D.S. Rankin129, N.P. Rapheeha34j, H. Rasheed28b, A. Rastogi18a, S. Rave100, S. Ravera56b,56a, B. Ravina37, I. Ravinovich171, M. Raymond37, A.L. Read126, N.P. Readioff142, D.M. Rebuzzi72a,72b

A.S. Reed58, K. Reeves27, D. Reikher37, T. Reisch55, A. Rej48, H. Ren61, M. Renda28b, F. Renner47, A.G. Rennie58, M. Repik55, A.L. Rescia43b,43a, S. Resconi70a, M. Ressegotti56b, S. Rettie116, W.F. Rettie35, M.M. Revering33, O.L. Rezanova38, P. Reznicek134, H. Riani36d, N. Ribaric37, B. Ricci68a,68c, R. Richter110, E. Richter-Was85b, M. Ridel128, S. Ridouani36d, P. Riedler37, E.M. Riefel46a,46b, J.O. Rieger116, M. Rimoldi34c, L. Rinaldi24b,24a, P. Rincke163,54, G. Ripellino163, I. Riu13, R. Riva77a,77b, J.C. Rivera Vergara167, F. Rizatdinova122, E. Rizvi94, B.R. Roberts39, S.S. Roberts137, D. Robinson33, A. Robson58, A. Rocchi75a,75b, C. Roda73a,73b, F.A. Rodriguez117, S. Rodriguez Bosca37, Y. Rodriguez Garcia23a, A.M. Rodríguez Vera117, S. Roe37, J.T. Roemer37, O. Røhne126, R.A. Rojas37, Z. Rokavec93, C.P.A. Roland128, A. Romaniouk78, E. Romano72a,72b, M. Romano24b, N. Rompotis92, L. Roos128, S. Rosati74a, L. Roscher47, B.J. Rosser39, E. Rossi127, E. Rossi71a,71b, L.P. Rossi60, L. Rossini53, R. Rosten120, M. Rotaru28b, R. Roth37, F.A. Rothen55, D. Rousseau65, D. Rousso47, S. Roy-Garand55, A. Rozanov102, Z.M.A. Rozario58, Y. Rozen153, A. Rubio Jimenez165, V.H. Ruelas Rivera19, T.A. Ruggeri1, A. Ruggiero127, A. Ruiz-Martinez165, A. Rummler37, G.B. Rupnik Boero37, N.A. Rusakovich38, S. Ruscelli48, H.L. Russell167, G. Russo137, J.P. Rutherfoord7, S. Rutherford Colmenares118, M. Rybar134, P. Rybczynski85a, A. Ryzhov44, M.A.E. Saadawy44, M. Sabate Gilarte37, F. Safai Tehrani74a, S. Saha1, B. Sahoo171, B.T. Saifuddin121, M. Saimpert1, I. Sain Saenz Diez62a, G.T. Saito81c, M. Saito156, T. Saitō156, A. Sala120, O.T. Salin65, A. Salnikov146, J. Salt165, A. Salvador Salas154, F. Salvatore149, G. Salvi106, A. Salzburger37, D. Sammel53, E. Sampson91, D. Sampsonidis155,d, D. Sampsonidou124, M.A.A. Samy58, J. Sánchez165, H. Sandaker126, C.O. Sander47, J.A. Sandesara172, M. Sandhoff173, C. Sandoval23b, L. Sanfilippo62a, D.P.C. Sankey135, T. Sano87, A. Sansar22c, A. Sansoni52, M. Santana Queiroz18b, L. Santi37, C. Santoni40, G. Santoro43b,43a, H. Santos131a,131b, M. Santos Aguiar97,81a, L. Santos Pereira Trigo47, E. Sanzani24b,24a, K.A. Saoucha83d, J.G. Saraiva131a,131d, J. Sardain7, S. Sarkar50, O. Sasaki82, K. Sato160, C. Sauer37, E. Sauvan4, P. Savard158,ak, M. Savic164, R. Sawada156, C. Sawyer97, A.M. Sayed27, C. Sbarra24b, A. Sbrizzi24b,24a, R. Scaglioni72a,72b, T. Scanlon96, J. Schaarschmidt140, U. Schäfer100, A.C. Schaffer65,44, D. Schaile109, R.D. Schamberger148, C. Scharf19, M.M. Schefer20, D. Scheirich134, M. Schernau138f, C. Scheulen55, C. Schiavi56b,56a, M. Schioppa43b,43a, S. Schlenker37, T. Schlomer54, J. Schmeing173, C.R. Schmidt49, E. Schmidt110, K. Schmieden25, C. Schmitt100, N. Schmitt100, S. Schmitt47, N.A. Schneider109, L. Schoeffel136, A. Schoening62b, P.G. Scholer35, E. Schopf144, M. Schott5, S. Schramm55, T. Schroer55, H-C. Schultz-Coulon23, M. Schumacher53, B.A. Schumm1, Ph. Schune55, H.R. Schwartz7, A. Schwarzman146, T.A. Schwarz106, Ph. Schwemling136, R. Schwienhorst107, F.G. Sciacca20, A. Sciandra30, G. Sciolla27, S.A. Scoville130, F. Scuri73a, C.D. Sebastiani37, K. Sedlaczek117, A. Sehrawat138b, S.C. Seidel114, B.D. Seidlitz41, C. Seitz47 J.M. Seixas81b,*, G. Sekhniaidze71a, L. Selem128, N. Semprini-Cesari24b,24a, A. Semushin175, V. Senthilkumar116, L. Serin65, M. Sessa71a,71b, H. Severini121, F. Sforza56b,56a, A. Sfyrla55, H. Shaddix117, A.H. Shah33, J.D. Shahinian129, M. Shamim37, L.Y. Shan14, M. Shapiro18a, A. Sharma37, A.S. Sharma166, P. Sharma30, K. Shaw149, S.M. Shaw101, D. Shemyakin171, Q. Shen14, D.J. Sheppard145, P. Sherwood96, L. Shi112b, X. Shi14, E.B. Shields171, S. Shimizu82, S. Shirabe88, M. Shiyakova38,z, M.J. Shochet39, D.R. Shope126, S. Shrestha120,ar, I. Shreyber38, M.J. Shroff104, P. Sicho132, A.M. Sickles164, E. Sideras Haddad34j, A.C. Sidley116, A. Sidoti24b, F. Siegert49, Dj. Sijacki16, F. Sili61, J.M. Silva47, I. Silva Ferreira81b, M.V. Silva Oliveira30, S.B. Silverstein46a, S. Simion65, R. Simoniello37, E.L. Simpson101, H. Simpson149, L.R. Simpson6, S. Simsek80, S.N. Singh27, S. Singh30, S. Sinha47, S. Sinha101, M. Sioli24b,24a, K. Sioulas9, E. Sitnikova47, J. Sjölin46a,46b, T.B. Sjursen17, A. Skaf54, E. Skorda21, P. Skubic121, M. Slawinska86, I. Slazyk17, I. Sliusar126, V. Smakhtin171, B.H. Smart135, Y. Smirnov34c, O. Smirnova98, J.L. Smith101, M.B. Smith35, R. Smith146, H. Smitmanns100, M. Smizanska91, K. Smolek133, P. Smolyanskiy133, A.A. Snesarev38, H.L. Snoek116, R.M. Snyder50, S. Snyder30, R. Sobie167,ab, A. Soffer154, C.A. Solans Sanchez37, M. Soldani37, E.Yu. Soldatov38, U. Soldevila165, A.A. Solodkov34j, S. Solomon27, A. Soloshenko38, O.V. Solovyanov40, P. Sommer49, A. Sopczak133, A.L. Sopio51, F. Sopkova29b, J.D. Sorenson114, I.R. Sotarriva Alvarez139, V. Sothilingam62a, O.J. Soto Sandoval138c,138b, S. Sottocornola67, R. Soualah83a, D. South47, N. Soybelman171, S. Spagnolo69a,69b, A.S. Spellman124, D. Sperlich53, B. Spisso71a,71b, L. Splendori102, M. Spousta134, E.J. Staats35, R. Stamen62a, E. Stanecka86, W. Stanek-Maslouska47, B. Stanislaus18a, M.M. Stanitzki47, G.H. Stark137, J. Stark89, P. Staroba132, P. Starovoitov83d, R. Staszewski86, C. Stauch109, G. Stavropoulos45, A. Stefl37, A. Stein100, P. Steinberg30, B. Stelzer145,159a, H.J. Stelzer130, O. Stelzer159a, H. Stenzel57, T.J. Stevenson149, G.A. Stewart47, G. Stoicea28b, M. Stolarski131a, S. Stonjek110, A. Straessner49, J. Strandberg147, S. Strandberg46a,46b, M. Stratmann173, M. Strauss121, T. Strebler102, P. Strizenec29b, R. Ströhmer168, D.M. Strom124, R. Stroynowski44, A. Strubig46a,46b, S.A. Stucci30, B. Stugu17, J. Stupak121, N.A. Styles47, D. Su146, S. Su61, X. Su61, D. Suchy29a, A.D. Sudhakar Ponnu54, L. Sudit171, Y. Sue82, K. Sugizaki129, D.M.S. Sultan127, L. Sultanaliyeva25, S. Sultansoy3b, S. Sun172, W. Sun14, S. Sundar Raman166, N. Sur98, J.P. Surdutovich120, N. Suri Jr174, M.R. Sutton149, M. Svatos132, P.N. Swallow33, S.N. Swatman37, M. Swiatlowski159a, A. Swoboda37, I. Sykora29a, M. Sykora134, T. Sykora134, D. Ta100, K. Tackmann47, A. Taffard162, R. Tafirout159a, Y. Takubo9, M. Talby102, N.M. Tamir13, A. Tanaka156, J. Tanaka156, R. Tanaka65, M. Tanasini148, Z. Tao166, S. Tapia Araya138g, S. Tappapoge100, A. Tarek Abouelfadl Mohamed37, S. Tarem153, K. Tariq14, G. Tarna37, G.F. Tartarelli70a, M.J. Tartarin141b,141a, P. Tas134, M. Tasevsky132, E. Tassi43b,43a, Y. Tayalati36e,aa, G.N. Taylor105, W. Taylor159b, R.J. Taylor Vara165, A.S. Tegetmeier89, P. Teixeira-Dias95, J.J. Teoh158, K. Terashi156, J. Terron99, S. Terzo13, M. Testa52, R.J. Teuscher158,ab, A. Thaler78, T. Thevenaux-Pelzer102, J.P. Thomas21, E.A. Thompson18a, P.D. Thompson21, E. Thomson129, R.E. Thornberry30, T.M. Thory-Rao21, C.N. Thotamuna Wijewardhana148, C. Tian61, V. Tikhomirov80, Yu.A. Tikhonov38, D. Timoshyn134, E.X.L. Ting1, P. Tipton174, A. Tishchman-Charny139, K. Todome134, S. Todorova-Nova134, L. Toffolin68a,68c, M. Togawa82, J. Tojo88, S. Tokár29a, O. Toldaeiev67, A.J. Toler103, G. Tolochek102, M. Tonomura82, L. Tompkins146, E. Torrence124, H. Torres89, D.I. Torres Arza54, E. Torres Reoyo165, E. Torró Pastor165, M. Toscani31, C. Tosciri39, M. Tost11, D.R. Tovey142, T. Trefzger168, A … P.M. Tricarico ¹³, A. Tricoli ³⁰, I.M. Trigger ¹⁵⁹ᵃ, S. Trincaz-Duvoid ¹²⁸, D.A. Trischuk ¹⁶⁷, A. Tropina³⁸, D. Truncali ⁷⁵ᵃ,⁷⁵ᵇ, L. Truong ³⁴ᶜ, M. Trzebinski ⁸⁶, A. Trzupek ⁸⁶, F. Tsai ¹⁴⁸, A. Tsiamis ¹⁵⁵, P.V. Tsiareshka³⁸, S. Tsigaridas ¹⁵⁹ᵃ, A. Tsirigotis ¹⁵⁵,ᵗ, E.G. Tskhadadze ¹⁵²ᵃ, H.F. Tsoi ¹²⁹, Y. Tsujikawa ⁸⁷, V. Tsulaia ¹⁸ᵃ, K. Tsuri ¹¹⁹, D. Tsybychev ¹⁴⁸, Y. Tu ⁶³ᵇ, A. Tudorache ²⁸ᵇ, V. Tudorache ²⁸ᵇ, S.B. Tuncay ¹²⁷, S. Turchikhin ⁵⁶ᵇ,⁵⁶ᵃ, I. Turk Cakir ³ᵃ, R. Turra ⁷⁰ᵃ, T. Turtuvshin ³⁸,ᵃᶜ, P.M. Tuts ⁴¹, Y. Uematsu ⁸², F. Ukegawa ¹⁶⁰, P.A. Ulloa Poblete ¹³⁸ᶜ,¹³⁸ᵇ, G. Unal ³⁷, A. Undrus ³⁰, J. Urban ²⁹ᵇ, P. Urrejola ¹³⁸ᵉ, G. Usai ⁸, R. Ushioda ¹⁵⁷, M. Usman ¹⁰⁸, F. Ustuner ⁵¹, Z. Uysal ⁸⁰, V. Vacek ¹³³, B. Vachon ¹⁰⁴, A. Vaitkus ⁹⁶, C. Valderanis ¹⁰⁹, E. Valdes Santurio ⁴⁶ᵃ,⁴⁶ᵇ, M. Valente ³⁷, S. Valentinetti ²⁴ᵇ,²⁴ᵃ, A. Valero ¹⁶⁵, E. Valiente Moreno ¹⁶⁵, S. Valjee ⁹⁴, A. Vallier ⁸⁹, J.A. Valls Ferrer ¹⁶⁵, D.R. Van Arneman ¹¹⁶, R. Van Den Broucke ¹²⁸, H.Z. Van Der Schyf ³⁴ʲ, P. Van Gemmeren ⁶, M. Van Rijnbach ³⁷, S. Van Stroud ⁹⁶, I. Van Vulpen ¹¹⁶, P. Vana ¹³⁴, M. Vanadia ⁷⁵ᵃ,⁷⁵ᵇ, U.M. Vande Voorde ¹⁴⁷, W. Vandelli ³⁷, E.R. Vandewall ¹⁴⁶, D. Vannicola ¹⁵⁴, R. Vari ⁷⁴ᵃ, M. Varma ¹⁷⁴, E.W. Varnes ⁷, C. Varni ⁸⁵ᵃ, D. Varouchas ⁶⁵, K.E. Varvell ¹⁵⁰, M.E. Vasile ²⁸ᵇ, A. Vasileiadou⁹, L. Vaslin⁸², M.D. Vassilev ¹⁴⁶, A. Vasyukov ³⁸, L.M. Vaughan ⁶, R. Vavricka¹³⁴, T. Vazquez Schroeder ¹³, J. Veatch ³², V. Vecchio ¹⁰¹, M.J. Veen ¹⁰³, I. Veliscek ³⁰, I. Velkovska ⁹³, L.M. Veloce ¹⁵⁸, F. Veloso ¹³¹ᵃ,¹³¹ᶜ, A.G. Veltman ⁵¹, S.H. Venetianer ¹⁶¹, S. Veneziano ⁷⁴ᵃ, A. Ventura ⁶⁹ᵃ,⁶⁹ᵇ, A. Verbytskyi ¹¹⁰, M. Verducci ⁷³ᵃ,⁷³ᵇ, C. Vergis ⁹⁴, M. Verissimo De Araujo ⁸¹ᵇ, W. Verkerke ¹¹⁶, J.C. Vermeulen ¹¹⁶, C. Vernieri ¹⁴⁶, M. Vessella ¹⁶², M.C. Vetterli ¹⁴⁵,ᵃᵏ, A. Vgenopoulos ¹⁰⁰, N. Viaux Maira ¹³⁸ᵍ,ᵃᶠ, L. Vicenik ¹³³, T. Vickey ¹⁴², O.E. Vickey Boeriu ¹⁴², G.H.A. Viehhauser ¹²⁷, L. Vigani ⁶²ᵇ, M. Vigl ¹¹⁰, M. Villa ²⁴ᵇ,²⁴ᵃ, M. Villaplana Perez ¹⁶⁵, E.M. Villhauer³⁹, E. Vilucchi ⁵², M. Vincent ¹⁶⁵, M.G. Vincter ³⁵, A. Visibile ⁴⁶ᵃ,⁴⁶ᵇ, A. Visive ¹¹⁶, C. Vittori ²⁴ᵇ,²⁴ᵃ, I. Vivarelli ²⁴ᵇ,²⁴ᵃ, M.I. Vivas Albornoz ⁴⁷, E. Voevodina ¹¹⁰, F. Vogel ¹⁰⁹, J.C. Voigt ¹³⁶, P. Vokac ¹³³, Yu. Volkotrub ⁸⁵ᵇ, L. Vomberg ²⁵, E. Von Toerne ²⁵, B. Vormwald ³⁷, K. Vorobev ⁵⁰, M. Vos ¹⁶⁵, K. Voss ⁴⁶ᵃ,⁴⁶ᵇ, M. Vozak ³⁷, L. Vozdecky ¹²¹, N. Vranjes ¹⁶, M. Vranjes Milosavljevic ¹⁶, M. Vreeswijk ¹¹⁶, N.K. Vu ¹¹²ᵃ, R. Vuillermet ³⁷, I. Vukotic ³⁹, I.K. Vyas ³⁵, J.F. Wack ³³, A. Wada ¹¹¹, S. Wada ¹⁶⁰, C. Wagner ¹⁴⁶, J.M. Wagner ¹⁸ᵃ, W. Wagner ¹⁷³, S. Wahdan ¹⁷³, H. Wahlberg ⁹⁰, C.H. Waits ¹²¹, R. Walker ¹⁰⁹, K. Walkingshaw Pass ⁵⁸, W. Walkowiak ¹⁴⁴, A. Wall ¹²⁹, E.J. Wallin ⁹⁸, T. Wamorkar ¹⁴⁶, K. Wandall-Christensen ¹⁶⁵, A. Wang ⁶¹, A.Z. Wang ¹³⁷, C. Wang ⁴⁷, C. Wang ¹¹, H. Wang ¹⁸ᵃ, J. Wang ⁶³ᶜ, P. Wang ¹⁰¹, P. Wang ⁹⁶, R. Wang ⁶⁰, R. Wang ¹⁰⁶, R. Wang ⁶, S.M. Wang ¹⁵¹, S. Wang ¹⁴,ᵃᵒ, T. Wang ¹¹⁵, T. Wang ⁶¹, W.T. Wang ¹²⁷, W. Wang ¹¹³ᶜ, X. Wang ¹⁶⁴, X. Wang ¹⁴¹ᵃ, X. Wang ⁴⁷, Y. Wang ¹⁴⁸, Y. Wang ¹¹⁴, Y. Wang ⁶¹, Z. Wang ⁴, Z. Wang⁶³ᵇ, C. Wanotayaroj ⁸², A. Warburton ¹⁰⁴, A.L. Warnerbring ¹⁴⁴, S. Waterhouse ⁹⁶, A.T. Watson ²¹, H. Watson ⁵¹, M.F. Watson ²¹, E. Watton ³⁷, G. Watts ¹⁴⁰, B.M. Waugh ⁹⁶, J.M. Webb ⁵³, C. Weber ³⁰, M.S. Weber ²⁰, C. Wei ⁶¹, Y. Wei ⁵³, A.R. Weidberg ¹²⁷, E.J. Weik ¹¹⁸, J. Weingarten ⁴⁸, C. Weiser ⁵³, P.S. Wells ³⁷, T. Wenaus ³⁰, T. Wengler ³⁷, N.S. Wenke¹¹⁰, N. Wermes ²⁵, D. Werner ⁴⁷, M. Wessels ⁶²ᵃ, A.M. Wharton ⁹¹, A.S. White ³⁷, A. White ⁸, M.J. White ¹, D. Whiteson ¹⁶², W. Wiedenmann ¹⁷², M. Wielers ¹³⁵, R. Wierda ¹⁴⁷, C. Wiglesworth ⁴², H.G. Wilkens ³⁷, J.J.H. Wilkinson ³³, S. Williams ³³, S. Willocq ¹⁰³, L.F. Wills¹¹⁷, D.J. Wilson ¹⁰¹, P.J. Windischhofer ³⁹, F.I. Winkel ³¹, F. Winklmeier ¹²⁴, B.T. Winter ⁵³, M. Wittgen ¹⁴⁶, M. Wobisch ⁹⁷, T. Wojtkowski⁵⁹, Z. Wolffs ¹¹⁶, J. Wollrath³⁷, M.W. Wolter ⁸⁶, H. Wolters ¹⁶³,¹³¹ᵃ,¹³¹ᶜ, M.C. Wong ¹³⁷, E.L. Woodward ⁵⁴, S.D. Worm ⁴⁷, B.K. Wosiek ⁸⁶, K.A. Wozniak ⁵⁵, W. Woźniak ⁸⁶, S. Wozniak ⁵⁴, K. Wraight ⁵⁸, C. Wu ¹⁵⁸, J. Wu ¹⁵⁶, M. Wu ¹¹²ᵇ, M. Wu ¹¹⁵, S.L. Wu ¹⁷², S. Wu ¹⁴,ᵃᵒ, X. Wu ⁶¹, Y.Q. Wu ¹⁵⁸, Y. Wu ⁶¹, Z. Wu ¹⁰², Z. Wu ¹¹²ᵃ, J. Wuerzinger ¹¹⁰, T.R. Wyatt ¹⁰¹, B.M. Wynne 51, L. Xia 112a, M. Xie 61, I. Xiotidis 37, E. Xochelli 37, D. Xu 14, H. Xu 61, L. Xu 61, T. Xu 106, W. Xu 112a, Y. Xu 140, Z. Xu 51, R. Xue 130, B. Yabsley 150, S. Yacoob 11, Y. Yamaguchi 82, E. Yamashita 156, H. Yamauchi 160, T. Yamazaki 18a, Y. Yamazaki 84, F. Yan 2, S. Yan 58, Z. Yan 103, C. Yang 18a, H.J. Yang 141a, H.T. Yang 61, S. Yang 61, X. Yang 37, X. Yang 14, Y. Yang 156, Y. Yang 61, W-M. Yao 18a, C.L. Yardley 149, J. Ye 14, S. Ye 30, X. Ye 106, I. Yeletskikh 38, B. Yeo 18b, M.R. Yexley 96, T.P. Yildirim 127, K. Yorita 156, C.J.S. Young 37, C. Young 146, I.N.L. Young 58, N.D. Young 124, D. Yu 141b,5, Y. Yu 61, J. Yuan 14,112c,ao, M. Yuan 106, R. Yuan 141b, M. Zaazoua 61, B. Zabinski 86, I. Zahir 36a, Q.U.A. Zahoor 51, A. Zaio 56b,56a, Z.K. Zak 86, T. Zakareishvili 165, S. Zambito 55, J. Zang 156, R. Zanzottera 70a,70b, O. Zaplatilek 133, I. Zatocilova 53, E. Zaya 147, C. Zeitnitz 173, H. Zeng 14, T. Ženiš 29a, S. Zenz 94, W. Zhan 61, B. Zhang 169, D.F. Zhang 142, G. Zhang 14,ao, J. Zhang 113b, J. Zhang 6, L. Zhang 61, L. Zhang 112a, P. Zhang 112a,112c, R. Zhang 112a, S. Zhang 14, T. Zhang 14, Y. Zhang 140, Y. Zhang 61, Y. Zhang 112a, Y. Zhang 15, Z. Zhang 149, Z. Zhang 101, Z. Zhang 18a, Z. Zhang 113b, Z. Zhang 65, H. Zhao 140, T. Zhao 113b, Y. Zhao 35, Z. Zhao 61, Z. Zhao 61, A. Zhemchugov 38, J. Zheng 112a, L. Zheng 113b, X. Zheng 61, Z. Zheng 146, D. Zhong 164, B. Zhou 106, B. Zhou 141b,141a, N. Zhou 141a, Y. Zhou 15, Y. Zhou 112a, Z. Zhou 61, J. Zhu 106, X. Zhu 141b, Y. Zhu 141a, X. Zhuang 14, K. Zhukov 67, P. Ziakas 4, N.I. Zimine 38, J. Zinsser 62b, M. Ziolkowski 144, L. Živković 16, A. Zoccoli 24b,24a, K. Zoch 37, A. Zografos 37, T.G. Zorbas 142, O. Zormpa 37, D. Zubov 155, L. Zwalinski 37.

1Department of Physics, University of Adelaide, Adelaide; Australia.

2Department of Physics, University of Alberta, Edmonton AB; Canada.

3(a)Department of Physics, Ankara University, Ankara;(b)Division of Physics, TOBB University of Economics and Technology, Ankara; Türkiye.

4LAPP, Université Savoie Mont Blanc, CNRS/IN2P3, Annecy; France.

5APC, Université Paris Cité, CNRS/IN2P3, Paris; France.

6High Energy Physics Division, Argonne National Laboratory, Argonne IL; United States of America.

7Department of Physics, University of Arizona, Tucson AZ; United States of America.

8Department of Physics, University of Texas at Arlington, Arlington TX; United States of America.

9Physics Department, National and Kapodistrian University of Athens, Athens; Greece.

10Physics Department, National Technical University of Athens, Zografou; Greece.

11Department of Physics, University of Texas at Austin, Austin TX; United States of America.

12Institute of Physics, Azerbaijan Academy of Sciences, Baku; Azerbaijan.

13Institut de Física d’Altes Energies (IFAE), Barcelona Institute of Science and Technology, Barcelona; Spain.

14Institute of High Energy Physics, Chinese Academy of Sciences, Beijing; China.

15Physics Department, Tsinghua University, Beijing; China.

16Institute of Physics, University of Belgrade, Belgrade; Serbia.

17Department for Physics and Technology, University of Bergen, Bergen; Norway.

18(a)Physics Division, Lawrence Berkeley National Laboratory, Berkeley CA;(b)University of California, Berkeley CA; United States of America.

19Institut für Physik, Humboldt Universität zu Berlin, Berlin; Germany.

20Albert Einstein Center for Fundamental Physics and Laboratory for High Energy Physics, University of Bern, Bern; Switzerland.

21School of Physics and Astronomy, University of Birmingham, Birmingham; United Kingdom.

22(a)Department of Physics, Bogazici University, Istanbul;(b)Department of Physics Engineering, Gaziantep University, Gaziantep;(c)Department of Physics, Istanbul University, Istanbul; Türkiye. 23 (a) Facultad de Ciencias y Centro de Investigaciones, Universidad Antonio Nariño, Bogotá; (b) Departamento de Física, Universidad Nacional de Colombia, Bogotá; Colombia.

24 (a) Dipartimento di Fisica e Astronomia A. Righi, Università di Bologna, Bologna; (b) INFN Sezione di Bologna; Italy.

25 Physikalisches Institut, Universität Bonn, Bonn; Germany.

26 Department of Physics, Boston University, Boston MA; United States of America.

27 Department of Physics, Brandeis University, Waltham MA; United States of America.

28 (a) Transilvania University of Brasov, Brasov; (b) Horia Hulubei National Institute of Physics and Nuclear Engineering, Bucharest; (c) Department of Physics, Alexandru Ioan Cuza University of Iasi, Iasi; (d) National Institute for Research and Development of Isotopic and Molecular Technologies, Physics Department, Cluj-Napoca; (e) National University of Science and Technology Politechnica, Bucharest; (f) West University in Timisoara, Timisoara; (g) Faculty of Physics, University of Bucharest, Bucharest; Romania.

29 (a) Faculty of Mathematics, Physics and Informatics, Comenius University, Bratislava; (b) Department of Subnuclear Physics, Institute of Experimental Physics of the Slovak Academy of Sciences, Kosice; Slovak Republic.

30 Physics Department, Brookhaven National Laboratory, Upton NY; United States of America.

31 Universidad de Buenos Aires, Facultad de Ciencias Exactas y Naturales, Departamento de Física, y CONICET, Instituto de Física de Buenos Aires (IFIBA), Buenos Aires; Argentina.

32 California State University, CA; United States of America.

33 Cavendish Laboratory, University of Cambridge, Cambridge; United Kingdom.

34 (a) Department of Physics, University of Cape Town, Cape Town; (b) iThemba Labs, Western Cape; (c) Department of Mechanical Engineering Science, University of Johannesburg, Johannesburg; (d) National Institute of Physics, University of the Philippines Diliman (Philippines); (e) Department of Physics, Stellenbosch University, Matieland; (f) University of KwaZulu-Natal, School of Agriculture and Science, Mathematics, Westville; (g) University of South Africa, Department of Physics, Pretoria; (h) University of Pretoria, Department of Mechanical and Aeronautical Engineering, Pretoria; (i) University of Zululand, KwaDlangezwa; (j) School of Physics, University of the Witwatersrand, Johannesburg; South Africa.

35 Department of Physics, Carleton University, Ottawa ON; Canada.

36 (a) Faculté des Sciences Ain Chock, Université Hassan II de Casablanca; (b) Faculté des Sciences, Université Ibn-Tofail, Kénitra; (c) Faculté des Sciences Semlalia, Université Cadi Ayyad, LPHEA-Marrakech; (d) LPMR, Faculté des Sciences, Université Mohamed Premier, Oujda; (e) Faculté des sciences, Université Mohammed V, Rabat; (f) Institute of Applied Physics, Mohammed VI Polytechnic University, Ben Guerir; Morocco.

37 CERN, Geneva; Switzerland.

38 Affiliated with an international laboratory covered by a cooperation agreement with CERN.

39 Enrico Fermi Institute, University of Chicago, Chicago IL; United States of America.

40 LPC, Université Clermont Auvergne, CNRS/IN2P3, Clermont-Ferrand; France.

41 Nevis Laboratory, Columbia University, Irvington NY; United States of America.

42 Niels Bohr Institute, University of Copenhagen, Copenhagen; Denmark.

43 (a) Dipartimento di Fisica, Università della Calabria, Rende; (b) INFN Gruppo Collegato di Cosenza, Laboratori Nazionali di Frascati; Italy.

44 Physics Department, Southern Methodist University, Dallas TX; United States of America.

45 National Centre for Scientific Research "Demokritos", Agia Paraskevi; Greece.

46 (a) Department of Physics, Stockholm University; (b) Oskar Klein Centre, Stockholm; Sweden.

47 Deutsches Elektronen-Synchrotron DESY, Hamburg and Zeuthen; Germany.

48 Fakultät Physik, Technische Universität Dortmund, Dortmund; Germany. 49Institut für Kern- und Teilchenphysik, Technische Universität Dresden, Dresden; Germany.

50Department of Physics, Duke University, Durham NC; United States of America.

51SUPA - School of Physics and Astronomy, University of Edinburgh, Edinburgh; United Kingdom.

52INFN e Laboratori Nazionali di Frascati, Frascati; Italy.

53Physikalisches Institut, Albert-Ludwigs-Universität Freiburg, Freiburg; Germany.

54II. Physikalisches Institut, Georg-August-Universität Göttingen, Göttingen; Germany.

55Département de Physique Nucléaire et Corpusculaire, Université de Genève, Genève; Switzerland.

56(a)Dipartimento di Fisica, Università di Genova, Genova;(b) INFN Sezione di Genova; Italy.

57II. Physikalisches Institut, Justus-Liebig-Universität Giessen, Giessen; Germany.

58SUPA - School of Physics and Astronomy, University of Glasgow, Glasgow; United Kingdom.

59LPSC, Université Grenoble Alpes, CNRS/IN2P3, Grenoble INP, Grenoble; France.

60Laboratory for Particle Physics and Cosmology, Harvard University, Cambridge MA; United States of America.

61Department of Modern Physics and State Key Laboratory of Particle Detection and Electronics, University of Science and Technology of China, Hefei; China.

62(a)Kirchhoff-Institut für Physik, Ruprecht-Karls-Universität Heidelberg, Heidelberg;(b)Physikalisches Institut, Ruprecht-Karls-Universität Heidelberg, Heidelberg; Germany.

63(a)Department of Physics, Chinese University of Hong Kong, Shatin, N.T., Hong Kong;(b)Department of Physics, University of Hong Kong, Hong Kong;(c)Department of Physics and Institute for Advanced Study, Hong Kong University of Science and Technology, Clear Water Bay, Kowloon, Hong Kong; China.

64Department of Physics, National Tsing Hua University, Hsinchu; Taiwan.

65IJCLab, Université Paris-Saclay, CNRS/IN2P3, 91405, Orsay; France.

66Centro Nacional de Microelectrónica (IMB-CNM-CSIC), Barcelona; Spain.

67Department of Physics, Indiana University, Bloomington IN; United States of America.

68(a)INFN Gruppo Collegato di Udine, Sezione di Trieste, Udine;(b)ICTP, Trieste;(c)Dipartimento Politecnico di Ingegneria e Architettura, Università di Udine, Udine; Italy.

69(a)INFN Sezione di Lecce;(b)Dipartimento di Matematica e Fisica, Università del Salento, Lecce; Italy.

70(a)INFN Sezione di Milano;(b)Dipartimento di Fisica, Università di Milano, Milano; Italy.

71(a)INFN Sezione di Napoli;(b)Dipartimento di Fisica, Università di Napoli, Napoli; Italy.

72(a)INFN Sezione di Pavia;(b)Dipartimento di Fisica, Università di Pavia, Pavia; Italy.

73(a)INFN Sezione di Pisa;(b)Dipartimento di Fisica E. Fermi, Università di Pisa, Pisa; Italy.

74(a)INFN Sezione di Roma;(b)Dipartimento di Fisica, Sapienza Università di Roma, Roma; Italy.

75(a)INFN Sezione di Roma Tor Vergata;(b)Dipartimento di Fisica, Università di Roma Tor Vergata, Roma; Italy.

76(a)INFN Sezione di Roma Tre;(b)Dipartimento di Matematica e Fisica, Università Roma Tre, Roma; Italy.

77(a)INFN-TIFPA;(b)Università degli Studi di Trento, Trento; Italy.

78Universität Innsbruck, Department of Astro and Particle Physics, Innsbruck; Austria.

79Department of Physics and Astronomy, Iowa State University, Ames IA; United States of America.

80Istinye University, Sarıyer, İstanbul; Türkiye.

81(a)Departamento de Engenharia Elétrica, Universidade Federal de Juiz de Fora (UFJF), Juiz de Fora;(b)Universidade Federal do Rio De Janeiro COPPE/EE/IF, Rio de Janeiro;(c)Instituto de Física, Universidade de São Paulo, São Paulo;(d)Rio de Janeiro State University, Rio de Janeiro;(e)Federal University of Bahia, Bahia; Brazil.

82KEK, High Energy Accelerator Research Organization, Tsukuba; Japan.

83(a)Khalifa University of Science and Technology, Abu Dhabi;(b)New York University Abu Dhabi, Abu Dhabi;(c)United Arab Emirates University, Al Ain;(d)University of Sharjah, Sharjah; United Arab Emirates.

{sup}84`Graduate School of Science, Kobe University, Kobe; Japan.

{sup}85({sup}a) AGH University of Krakow, Faculty of Physics and Applied Computer Science, Krakow;({sup}b`) Marian Smoluchowski Institute of Physics, Jagiellonian University, Krakow; Poland.

{sup}86`Institute of Nuclear Physics Polish Academy of Sciences, Krakow; Poland.

{sup}87`Faculty of Science, Kyoto University, Kyoto; Japan.

{sup}88`Research Center for Advanced Particle Physics and Department of Physics, Kyushu University, Fukuoka ; Japan.

{sup}89`L2IT, Université de Toulouse, CNRS/IN2P3, UPS, Toulouse; France.

{sup}90`Instituto de Física La Plata, Universidad Nacional de La Plata and CONICET, La Plata; Argentina.

{sup}91`Physics Department, Lancaster University, Lancaster; United Kingdom.

{sup}92`Oliver Lodge Laboratory, University of Liverpool, Liverpool; United Kingdom.

{sup}93`Department of Experimental Particle Physics, Jožef Stefan Institute and Department of Physics, University of Ljubljana, Ljubljana; Slovenia.

{sup}94`Department of Physics and Astronomy, Queen Mary University of London, London; United Kingdom.

{sup}95`Department of Physics, Royal Holloway University of London, Egham; United Kingdom.

{sup}96`Department of Physics and Astronomy, University College London, London; United Kingdom.

{sup}97`Louisiana Tech University, Ruston LA; United States of America.

{sup}98`Fysiska institutionen, Lunds universitet, Lund; Sweden.

{sup}99`Departamento de Física Teorica C-15 and CIAFF, Universidad Autónoma de Madrid, Madrid; Spain.

{sup}100`Institut für Physik, Universität Mainz, Mainz; Germany.

{sup}101`School of Physics and Astronomy, University of Manchester, Manchester; United Kingdom.

{sup}102`CPPM, Aix-Marseille Université, CNRS/IN2P3, Marseille; France.

{sup}103`Department of Physics, University of Massachusetts, Amherst MA; United States of America.

{sup}104`Department of Physics, McGill University, Montreal QC; Canada.

{sup}105`School of Physics, University of Melbourne, Victoria; Australia.

{sup}106`Department of Physics, University of Michigan, Ann Arbor MI; United States of America.

{sup}107`Department of Physics and Astronomy, Michigan State University, East Lansing MI; United States of America.

{sup}108`Group of Particle Physics, University of Montreal, Montreal QC; Canada.

{sup}109`Fakultät für Physik, Ludwig-Maximilians-Universität München, München; Germany.

{sup}110`Max-Planck-Institut für Physik (Werner-Heisenberg-Institut), Garching; Germany.

{sup}111`Graduate School of Science and Kobayashi-Maskawa Institute, Nagoya University, Nagoya; Japan.

{sup}112({sup}a) Department of Physics, Nanjing University, Nanjing;({sup}b) School of Science, Shenzhen Campus of Sun Yat-sen University;({sup}c`) University of Chinese Academy of Science (UCAS), Beijing; China.

{sup}113({sup}a) School of Physics, Nankai University, Tianjin;({sup}b) Institute of Frontier and Interdisciplinary Science and Key Laboratory of Particle Physics and Particle Irradiation (MOE), Shandong University, Qingdao;({sup}c`) School of Physics, Zhengzhou University; China.

{sup}114`Department of Physics and Astronomy, University of New Mexico, Albuquerque NM; United States of America.

{sup}115`Institute for Mathematics, Astrophysics and Particle Physics, Radboud University/Nikhef, Nijmegen; Netherlands.

{sup}116`Nikhef National Institute for Subatomic Physics and University of Amsterdam, Amsterdam; Netherlands.

{sup}117`Department of Physics, Northern Illinois University, DeKalb IL; United States of America.

{sup}118`Department of Physics, New York University, New York NY; United States of America.

{sup}119`Ochanomizu University, Otsuka, Bunkyo-ku, Tokyo; Japan. 120Ohio State University, Columbus OH; United States of America.

121Homer L. Dodge Department of Physics and Astronomy, University of Oklahoma, Norman OK; United States of America.

122Department of Physics, Oklahoma State University, Stillwater OK; United States of America.

123Palacký University, Joint Laboratory of Optics, Olomouc; Czech Republic.

124Institute for Fundamental Science, University of Oregon, Eugene, OR; United States of America.

125Graduate School of Science, University of Osaka, Osaka; Japan.

126Department of Physics, University of Oslo, Oslo; Norway.

127Department of Physics, Oxford University, Oxford; United Kingdom.

128LPNHE, Sorbonne Université, Université Paris Cité, CNRS/IN2P3, Paris; France.

129Department of Physics, University of Pennsylvania, Philadelphia PA; United States of America.

130Department of Physics and Astronomy, University of Pittsburgh, Pittsburgh PA; United States of America.

131(a)Laboratório de Instrumentação e Física Experimental de Partículas - LIP, Lisboa;(b)Departamento de Física, Faculdade de Ciências, Universidade de Lisboa, Lisboa;(c)Departamento de Física, Universidade de Coimbra, Coimbra;(d)Centro de Física Nuclear da Universidade de Lisboa, Lisboa;(e)Departamento de Física, Escola de Ciências, Universidade do Minho, Braga;(f)Departamento de Física Teórica y del Cosmos, Universidad de Granada, Granada (Spain);(g)Departamento de Física, Instituto Superior Técnico, Universidade de Lisboa, Lisboa; Portugal.

132Institute of Physics of the Czech Academy of Sciences, Prague; Czech Republic.

133Czech Technical University in Prague, Prague; Czech Republic.

134Charles University, Faculty of Mathematics and Physics, Prague; Czech Republic.

135Particle Physics Department, Rutherford Appleton Laboratory, Didcot; United Kingdom.

136IRFU, CEA, Université Paris-Saclay, Gif-sur-Yvette; France.

137Santa Cruz Institute for Particle Physics, University of California Santa Cruz, Santa Cruz CA; United States of America.

138(a)Departamento de Física, Pontificia Universidad Católica de Chile, Santiago;(b)Millennium Institute for Subatomic physics at high energy frontier (SAPHIR), Santiago;(c)Instituto de Investigación Multidisciplinario en Ciencia y Tecnología, y Departamento de Física, Universidad de La Serena;(d)Universidad Andres Bello, Department of Physics, Santiago;(e)Universidad San Sebastian, Recoleta;(f)Instituto de Alta Investigación, Universidad de Tarapacá, Arica;(g)Departamento de Física, Universidad Técnica Federico Santa María, Valparaíso; Chile.

139Department of Physics, Institute of Science, Tokyo; Japan.

140Department of Physics, University of Washington, Seattle WA; United States of America.

141(a)State Key Laboratory of Dark Matter Physics, School of Physics and Astronomy, Shanghai Jiao Tong University, Key Laboratory for Particle Astrophysics and Cosmology (MOE), SKLPPC, Shanghai;(b)State Key Laboratory of Dark Matter Physics, Tsung-Dao Lee Institute, Shanghai Jiao Tong University, Shanghai; China.

142Department of Physics and Astronomy, University of Sheffield, Sheffield; United Kingdom.

143Department of Physics, Shinshu University, Nagano; Japan.

144Department Physik, Universität Siegen, Siegen; Germany.

145Department of Physics, Simon Fraser University, Burnaby BC; Canada.

146SLAC National Accelerator Laboratory, Stanford CA; United States of America.

147Department of Physics, Royal Institute of Technology, Stockholm; Sweden.

148Departments of Physics and Astronomy, Stony Brook University, Stony Brook NY; United States of America.

149Department of Physics and Astronomy, University of Sussex, Brighton; United Kingdom. {sup}150`School of Physics, University of Sydney, Sydney; Australia.

{sup}151`Institute of Physics, Academia Sinica, Taipei; Taiwan.

{sup}152{sup}(a)E. Andronikashvili Institute of Physics, Iv. Javakhishvili Tbilisi State University, Tbilisi;{sup}(b)High Energy Physics Institute, Tbilisi State University, Tbilisi;{sup}(c)University of Georgia, Tbilisi; Georgia.

{sup}153`Department of Physics, Technion, Israel Institute of Technology, Haifa; Israel.

{sup}154`Raymond and Beverly Sackler School of Physics and Astronomy, Tel Aviv University, Tel Aviv; Israel.

{sup}155`Department of Physics, Aristotle University of Thessaloniki, Thessaloniki; Greece.

{sup}156`International Center for Elementary Particle Physics and Department of Physics, University of Tokyo, Tokyo; Japan.

{sup}157`Graduate School of Science and Technology, Tokyo Metropolitan University, Tokyo; Japan.

{sup}158`Department of Physics, University of Toronto, Toronto ON; Canada.

{sup}159{sup}(a)TRIUMF, Vancouver BC;{sup}(b)`Department of Physics and Astronomy, York University, Toronto ON; Canada.

{sup}160`Division of Physics and Tomonaga Center for the History of the Universe, Faculty of Pure and Applied Sciences, University of Tsukuba, Tsukuba; Japan.

{sup}161`Department of Physics and Astronomy, Tufts University, Medford MA; United States of America.

{sup}162`Department of Physics and Astronomy, University of California Irvine, Irvine CA; United States of America.

{sup}163`Department of Physics and Astronomy, University of Uppsala, Uppsala; Sweden.

{sup}164`Department of Physics, University of Illinois, Urbana IL; United States of America.

{sup}165`Instituto de Física Corpuscular (IFIC), Centro Mixto Universidad de Valencia - CSIC, Valencia; Spain.

{sup}166`Department of Physics, University of British Columbia, Vancouver BC; Canada.

{sup}167`Department of Physics and Astronomy, University of Victoria, Victoria BC; Canada.

{sup}168`Fakultät für Physik und Astronomie, Julius-Maximilians-Universität Würzburg, Würzburg; Germany.

{sup}169`Department of Physics, University of Warwick, Coventry; United Kingdom.

{sup}170`Waseda University, Tokyo; Japan.

{sup}171`Department of Particle Physics and Astrophysics, Weizmann Institute of Science, Rehovot; Israel.

{sup}172`Department of Physics, University of Wisconsin, Madison WI; United States of America.

{sup}173`Fakultät für Mathematik und Naturwissenschaften, Fachgruppe Physik, Bergische Universität Wuppertal, Wuppertal; Germany.

{sup}174`Department of Physics, Yale University, New Haven CT; United States of America.

{sup}175`Yerevan Physics Institute, Yerevan; Armenia.

a Also at Affiliated with an institute formerly covered by a cooperation agreement with CERN.

b Also at An-Najah National University, Nablus; Palestine.

c Also at Borough of Manhattan Community College, City University of New York, New York NY; United States of America.

d Also at Center for Interdisciplinary Research and Innovation (CIRI-AUTH), Thessaloniki; Greece.

e Also at Centre of Physics of the Universities of Minho and Porto (CF-UM-UP); Portugal.

f Also at CERN, Geneva; Switzerland.

g Also at Département de Physique Nucléaire et Corpusculaire, Université de Genève, Genève; Switzerland.

h Also at Departament de Fisica de la Universitat Autonoma de Barcelona, Barcelona; Spain.

i Also at Department of Financial and Management Engineering, University of the Aegean, Chios; Greece.

j Also at Department of Modern Physics and State Key Laboratory of Particle Detection and Electronics, University of Science and Technology of China, Hefei; China.

k Also at Department of Physics, Ben Gurion University of the Negev, Beer Sheva; Israel.

l Also at Department of Physics, Bolu Abant Izzet Baysal University, Bolu; Türkiye.

{sup}m` Also at Department of Physics, King’s College London, London; United Kingdom.

{sup}n` Also at Department of Physics, Stellenbosch University; South Africa.

{sup}o` Also at Department of Physics, University of Fribourg, Fribourg; Switzerland.

{sup}p` Also at Department of Physics, University of Thessaly; Greece.

{sup}q` Also at Department of Physics, Westmont College, Santa Barbara; United States of America.

{sup}r` Also at Faculty of Physics, Sofia University, ’St. Kliment Ohridski’, Sofia; Bulgaria.

{sup}s` Also at Faculty of Physics, University of Bucharest; Romania.

{sup}t` Also at Hellenic Open University, Patras; Greece.

{sup}u` Also at Henan University; China.

{sup}v` Also at Imam Mohammad Ibn Saud Islamic University; Saudi Arabia.

{sup}w` Also at Indian Institute of Technology (IIT), Jodhpur; India.

{sup}x` Also at Institucio Catalana de Recerca i Estudis Avancats, ICREA, Barcelona; Spain.

{sup}y` Also at Institut für Experimentalphysik, Universität Hamburg, Hamburg; Germany.

{sup}z` Also at Institute for Nuclear Research and Nuclear Energy (INRNE) of the Bulgarian Academy of Sciences, Sofia; Bulgaria.

{sup}aa` Also at Institute of Applied Physics, Mohammed VI Polytechnic University, Ben Guerir; Morocco.

{sup}ab` Also at Institute of Particle Physics (IPP); Canada.

{sup}ac` Also at Institute of Physics and Technology, Mongolian Academy of Sciences, Ulaanbaatar; Mongolia.

{sup}ad` Also at Institute of Physics, Azerbaijan Academy of Sciences, Baku; Azerbaijan.

{sup}ae` Also at Institute of Theoretical Physics, Ilia State University, Tbilisi; Georgia.

{sup}af` Also at Millennium Institute for Subatomic physics at high energy frontier (SAPHIR), Santiago; Chile.

{sup}ag` Also at National Institute of Physics, University of the Philippines Diliman (Philippines); Philippines.

{sup}ah` Associated at Physics Division, Lawrence Berkeley National Laboratory, Berkeley CA; United States of America.

{sup}ai` Also at School of Physics, University of the Witwatersrand, Johannesburg; South Africa.

{sup}aj` Also at The Collaborative Innovation Center of Quantum Matter (CICQM), Beijing; China.

{sup}ak` Also at TRIUMF, Vancouver BC; Canada.

{sup}al` Also at Università di Napoli Parthenope, Napoli; Italy.

{sup}am` Also at Universita degli Studi Link; Italy.

{sup}an` Also at University and INFN Torino, Torino; Italy.

{sup}ao` Also at University of Chinese Academy of Sciences (UCAS), Beijing; China.

{sup}ap` Also at University of Colorado Boulder, Department of Physics, Colorado; United States of America.

{sup}aq` Also at University of Siena; Italy.

{sup}ar` Also at Washington College, Chestertown, MD; United States of America.

{sup}as` Also at Yeditepe University, Physics Department, Istanbul; Türkiye.

{sup}*` Deceased

References

  1. [1]ATLAS Collaboration, Measurement of the production cross section of jets in association with a Z boson in pp collisions at √s = 7 TeV with the ATLAS detector, JHEP 07 (2013) 032, arXiv: 1304.7098 [hep-ex].DOI
  2. [2]ATLAS Collaboration, Measurement of the kₜ splitting scales in Z → ℓℓ events in pp collisions at √s = 8 TeV with the ATLAS detector, JHEP 08 (2017) 026, arXiv: 1704.01530 [hep-ex].DOI
  3. [3]ATLAS Collaboration, Cross-section measurements for the production of a Z boson in association with high-transverse-momentum jets in pp collisions at √s = 13 TeV with the ATLAS detector, JHEP 06 (2023) 080, arXiv: 2205.02597 [hep-ex].DOI
  4. [4]ATLAS Collaboration, Simultaneous Unbinned Differential Cross-Section Measurement of Twenty-Four Z+jets Kinematic Observables with the ATLAS Detector, Phys. Rev. Lett. 133 (2024) 261803, arXiv: 2405.20041 [hep-ex].DOI
  5. [5]CMS Collaboration, Event shapes and azimuthal correlations in Z + jets events in pp collisions at √s = 7 TeV, Phys. Lett. B 722 (2013) 238, arXiv: 1301.1646 [hep-ex].DOI
  6. [6]CMS Collaboration, Rapidity distributions in exclusive Z + jet and γ + jet events in pp collisions at √s = 7 TeV, Phys. Rev. D 88 (2013) 112009, arXiv: 1310.3082 [hep-ex].DOI
  7. [7]CMS Collaboration, Measurements of jet multiplicity and differential production cross sections of Z + jets events in proton–proton collisions at √s = 7 TeV, Phys. Rev. D 91 (2015) 052008, arXiv: 1408.3104 [hep-ex].DOI
  8. [8]CMS Collaboration, Studies of jet mass in dijet and W/Z + jet events, JHEP 05 (2013) 090, arXiv: 1303.4811 [hep-ex].DOI
  9. [9]CMS Collaboration, Measurements of differential production cross sections for a Z boson in association with jets in pp collisions at √s = 8 TeV, JHEP 04 (2017) 022, arXiv: 1611.03844 [hep-ex].DOI
  10. [10]CMS Collaboration, Measurement of differential cross sections for Z boson production in association with jets in proton–proton collisions at √s = 13 TeV, Eur. Phys. J. C 78 (2018) 965, arXiv: 1804.05252 [hep-ex].DOI
  11. [11]CMS Collaboration, Measurements of the differential cross sections of the production of Z+jets and γ+jets and of Z boson emission collinear with a jet in pp collisions at √s = 13 TeV, JHEP 05 (2021) 285, arXiv: 2102.02238 [hep-ex].DOI
  12. [12]CMS Collaboration, Study of Z boson plus jets events using variables sensitive to double-parton scattering in pp collisions at 13 TeV, JHEP 10 (2021) 176, arXiv: 2105.14511 [hep-ex].DOI
  13. [13]CMS Collaboration, Measurement of differential cross sections for the production of a Z boson in association with jets in proton–proton collisions at √s = 13 TeV, Phys. Rev. D 108 (2023) 052004, arXiv: 2205.02872 [hep-ex].DOI
  14. [14]CMS Collaboration, Study of quark and gluon jet substructure in Z+jet and dijet events from pp collisions, JHEP 01 (2022) 188, arXiv: 2109.03340 [hep-ex].DOI
  15. [15]CMS Collaboration, Azimuthal correlations in Z+jets events in proton–proton collisions at √s = 13 TeV, Eur. Phys. J. C 83 (2023) 722, arXiv: 2210.16139 [hep-ex].DOI
  16. [16]LHCb Collaboration, Measurement of charged hadron production in Z-tagged jets in proton-proton collisions at √s = 8 TeV, Phys. Rev. Lett. 123 (2019) 232001, arXiv: 1904.08878 [hep-ex].
  17. [17]LHCb Collaboration, Multidifferential study of identified charged hadron distributions in Z-tagged jets in proton-proton collisions at √s = 13 TeV, Phys. Rev. D 108 (2023) L031103, arXiv: 2208.11691 [hep-ex].DOI
  18. [18]A. J. Larkoski, I. Moult, and B. Nachman, Jet Substructure at the Large Hadron Collider: A Review of Recent Advances in Theory and Machine Learning, Phys. Rept. 841 (2020) 1, arXiv: 1709.04464 [hep-ph].DOI
  19. [19]R. Kogler et al., Jet Substructure at the Large Hadron Collider, Rev. Mod. Phys. 91 (2019) 045003, arXiv: 1803.06991 [hep-ex].DOI
  20. [20]ATLAS Collaboration, Further ATLAS tunes of Pythia 6 and Pythia 8, ATL-PHYS-PUB-2011-014, 2011, URL: https://cds.cern.ch/record/1400677.
  21. [21]G. Cowan, A survey of unfolding methods for particle physics, Conf. Proc. C0203181 (2002) 248.
  22. [22]L. B. Lucy, An iterative technique for the rectification of observed distributions, Astron. J. 79 (1974) 745.DOI
  23. [23]W. H. Richardson, Bayesian-Based Iterative Method of Image Restoration*, J. Opt. Soc. Am. 62 (1972) 55.DOI
  24. [24]G. D’Agostini, A multidimensional unfolding method based on Bayes’ theorem, Nucl. Instrum. Meth. A 362 (1995) 487.
  25. [25]A. Höcker and V. Kartvelishvili, SVD approach to data unfolding, Nucl. Instrum. Meth. A 372 (1996) 469, arXiv: hep-ph/9509307.
  26. [26]S. Schmitt, TUnfold: an algorithm for correcting migration effects in high energy physics, JINST 7 (2012) T10003, arXiv: 1205.6201 [physics.data-an].DOI
  27. [27]A. Andreassen, P. T. Komiske, E. M. Metodiev, B. Nachman, and J. Thaler, OmniFold: A Method to Simultaneously Unfold All Observables, Phys. Rev. Lett. 124 (2020) 182001, arXiv: 1911.09107 [hep-ph].DOI
  28. [28]N. Gagunashvili, Machine learning approach to inverse problem and unfolding procedure, (2010), arXiv: 1004.2006 [physics.data-an].arxiv.org/abs/1004.2006
  29. [29]A. Glazov, Machine learning as an instrument for data unfolding, (2017), arXiv: 1712.01814 [physics.data-an].arxiv.org/abs/1712.01814
  30. [30]K. Datta, D. Kar, and D. Roy, Unfolding with Generative Adversarial Networks, (2018), arXiv: 1806.00433 [physics.data-an].arxiv.org/abs/1806.00433
  31. [31]M. Bellagente, A. Butter, G. Kasieczka, T. Plehn, and R. Winterhalder, How to GAN away Detector Effects, SciPost Phys. 8 (2020) 070, arXiv: 1912.00477 [hep-ph].DOI
  32. [32]M. Bellagente et al., Invertible networks or partons to detector and back again, SciPost Phys. 9 (2020) 074, URL: https://scipost.org/10.21468/SciPostPhys.9.5.074.
  33. [33]M. Vandegar, M. Kagan, A. Wehenkel, and G. Louppe, Neural Empirical Bayes: Source Distribution Estimation and its Applications to Simulation-Based Inference, (2021), arXiv: 2011.05836 [stat.ML].arxiv.org/abs/2011.05836
  34. [34]J. N. Howard, S. Mandt, D. Whiteson, and Y. Yang, Learning to simulate high energy particle collisions from unlabeled data, Sci. Rep. 12 (2022) 7567, arXiv: 2101.08944 [hep-ph].DOI
  35. [35]M. Backes, A. Butter, M. Dunford, and B. Malaescu, An unfolding method based on conditional invertible neural networks (cINN) using iterative training, SciPost Phys. Core 7 (2024) 007, arXiv: 2212.08674 [hep-ph].arxiv.org/abs/2212.08674
  36. [36]S. Diefenbacher, G.-H. Liu, V. Mikuni, B. Nachman, and W. Nie, Improving generative model-based unfolding with Schrödinger bridges, Phys. Rev. D 109 (2024) 076011, arXiv: 2308.12351 [hep-ph].arxiv.org/abs/2308.12351
  37. [37]A. Shmakov et al., End-To-End Latent Variational Diffusion Models for Inverse Problems in High Energy Physics, (2023), arXiv: 2305.10399 [hep-ex].arxiv.org/abs/2305.10399
  38. [38]J. Chan and B. Nachman, Unbinned profiled unfolding, Phys. Rev. D 108 (2023) 016002, arXiv: 2302.05390 [hep-ph].DOI
  39. [39]A. Shmakov et al., Full event particle-level unfolding with variable-length latent variational diffusion, SciPost Phys. 18 (2025) 117, arXiv: 2404.14332 [hep-ex].DOI
  40. [40]K. Desai, B. Nachman, and J. Thaler, Moment extraction using an unfolding protocol without binning, Phys. Rev. D 110 (2024) 116013, arXiv: 2407.11284 [hep-ph].
  41. [41]A. Butter et al., Generative Unfolding with Distribution Mapping, (2024), arXiv: 2411.02495 [hep-ph].DOI
  42. [42]H. Zhu et al., “Multidimensional Deconvolution with Profiling,” Advances in Neural Information Processing Systems, Machine Learning for the Physical Sciences Workshop, vol. 150, 2024, arXiv: 2409.10421 [hep-ph].arxiv.org/abs/2409.10421
  43. [43]C. Pazos et al., Towards universal unfolding of detector effects in high-energy physics using denoising diffusion probabilistic models, SciPost Phys. Core 8 (2025) 064, arXiv: 2406.01507 [physics.data-an].arxiv.org/abs/2406.01507
  44. [44]L. Favaro et al., How to Unfold Top Decays, (2025), arXiv: 2501.12363 [hep-ph].DOI
  45. [45]A. Butter, T. Heimel, N. Huetsch, M. Kagan, and T. Plehn, Simulation-Prior Independent Neural Unfolding Procedure, (2025), arXiv: 2507.15084 [hep-ph].arxiv.org/abs/2507.15084
  46. [46]A. Butter, N. Huetsch, V. Mikuni, B. Nachman, and S. Palacios Schweitzer, Analysis-ready Generative Unfolding, (2025), arXiv: 2509.02708 [hep-ph].arxiv.org/abs/2509.02708
  47. [47]A. Petitjean et al., Generative Unfolding of Jets and Their Substructure, (2025), arXiv: 2510.19906 [hep-ph].arxiv.org/abs/2510.19906
  48. [48]U. Sohail Qureshi, K. Desai, J. Thaler, and B. Nachman, Reweighting Adversarial Networks for Unbinned Unfolding, (2026), arXiv: 2606.06603 [hep-ph].arxiv.org/abs/2606.06603
  49. [49]A. Ore and T. Plehn, Unfolding without iterations, adversaries, or surrogates, SciPost Phys. 21 (2026) 055, URL: https://scipost.org/10.21468/SciPostPhys.21.3.055.
  50. [50]N. Huetsch et al., The landscape of unfolding with machine learning, SciPost Phys. 18 (2025) 070, arXiv: 2404.18807 [hep-ph], URL: https://scipost.org/10.21468/SciPostPhys.18.2.070.DOI
  51. [51]A. Andreassen et al., “Scaffolding Simulations with Deep Learning for High-dimensional Deconvolution,” 9th International Conference on Learning Representations, 2021, arXiv: 2105.04448 [stat.ML].arxiv.org/abs/2105.04448
  52. [52]H1 Collaboration, Measurement of Lepton-Jet Correlation in Deep-Inelastic Scattering with the H1 Detector Using Machine Learning for Unfolding, Phys. Rev. Lett. 128 (2022) 132002, arXiv: 2108.12376 [hep-ex].
  53. [53]H1 Collaboration, Unbinned deep learning jet substructure measurement in high Q2ep collisions at HERA, Phys. Lett. B 844 (2023) 138101, arXiv: 2303.13620 [hep-ex].
  54. [54]H1 Collaboration, Machine Learning-Assisted Measurement of Lepton-Jet Azimuthal Angular Asymmetries in Deep-Inelastic Scattering at HERA, (2024), arXiv: 2412.14092 [hep-ex].
  55. [55]STAR Collaboration, Measurement of CollinearDrop jet mass and its correlation with SoftDrop groomed jet substructure observables in √s = 200 GeV pp collisions by STAR, (2023), arXiv: 2307.07718 [nucl-ex].
  56. [56]T. Pani, Generalized angularities measurements from STAR at √sNN = 200 GeV, EPJ Web Conf. 296 (2024) 11003, arXiv: 2403.13921 [nucl-ex].arxiv.org/abs/2403.13921
  57. [57]R. G. Huang et al., Machine learning assisted unfolding for neutrino cross-section measurements with the OmniFold technique, Phys. Rev. D 112 (2025) 012008, arXiv: 2504.06857 [physics.data-an].DOI
  58. [58]ATLAS Collaboration, Measurement of jet track functions in pp collisions at √s = 13 TeV with the ATLAS detector, Phys. Lett. B 868 (2025) 139680, arXiv: 2502.02062 [hep-ex].DOI
  59. [59]CMS Collaboration, Measurement of event shapes in minimum-bias events from proton–proton collisions at √s = 13 TeV, Phys. Rev. D 112 (2025) 112006, arXiv: 2505.17850 [hep-ex].DOI
  60. [60]F. Canelli et al., A practical guide to unbinned unfolding, Eur. Phys. J. C 86 (2026) 106, arXiv: 2507.09582 [hep-ph].arxiv.org/abs/2507.09582
  61. [61]A. Altheimer et al., Boosted Objects and Jet Substructure at the LHC. Report of BOOST2012, held at IFIC Valencia, 23rd-27th of July 2012, Eur. Phys. J. C 74 (2014) 2792, arXiv: 1311.2708 [hep-ex].
  62. [62]ATLAS Collaboration, New techniques for jet calibration with the ATLAS detector, Eur. Phys. J. C 83 (2023) 761, arXiv: 2303.17312 [hep-ex].DOI
  63. [63]ATLAS Collaboration, Measurement of the Soft-Drop Jet Mass in pp Collisions at √s = 13 TeV with the ATLAS detector, Phys. Rev. Lett. 121 (2018) 092001, arXiv: 1711.08341 [hep-ex].DOI
  64. [64]ATLAS Collaboration, Accuracy versus precision in boosted top tagging with the ATLAS detector, JINST 19 (2024) P08018, arXiv: 2407.02127 [hep-ex].
  65. [65]ATLAS Collaboration, Unbinned measurement of Z+jets production, unfolded to full charged-particle phase space using the OmniFold algorithm, 2026, URL: https://opendata.cern/record/atlas-160005.
  66. [66]ATLAS Collaboration, Zjets Omnifold Public Code, 2026, URL: https://gitlab.cern.ch/atlas/open-data/sm-z-jets-omnifold-2026.
  67. [67]ATLAS Collaboration, The ATLAS Experiment at the CERN Large Hadron Collider, JINST 3 (2008) S08003.
  68. [68]ATLAS Collaboration, Luminosity determination in pp collisions at √s = 13 TeV using the ATLAS detector at the LHC, Eur. Phys. J. C 83 (2023) 982, arXiv: 2212.09379 [hep-ex].DOI
  69. [69]G. Avoni et al., The new LUCID-2 detector for luminosity measurement and monitoring in ATLAS, JINST 13 (2018) P07017.DOI
  70. [70]ATLAS Collaboration, Performance of the ATLAS trigger system in 2015, Eur. Phys. J. C 77 (2017) 317, arXiv: 1611.09661 [hep-ex].DOI
  71. [71]ATLAS Collaboration, Software and computing for Run 3 of the ATLAS experiment at the LHC, Eur. Phys. J. C 85 (2025) 234, arXiv: 2404.06335 [hep-ex], Erratum: Eur. Phys. J. C 85 (2025) 907.DOI
  72. [72]ATLAS Collaboration, Performance of the ATLAS muon triggers in Run 2, JINST 15 (2020) P09015, arXiv: 2004.13447 [physics.ins-det].DOI
  73. [73]ATLAS Collaboration, Muon reconstruction and identification efficiency in ATLAS using the full Run 2 pp collision data set at √s = 13 TeV, Eur. Phys. J. C 81 (2021) 578, arXiv: 2012.00578 [hep-ex].DOI
  74. [74]ATLAS Collaboration, Vertex Reconstruction Performance of the ATLAS Detector at √s = 13 TeV, ATL-PHYS-PUB-2015-026, 2015, URL: https://cds.cern.ch/record/2037717.
  75. [75]ATLAS Collaboration, Early Inner Detector Tracking Performance in the 2015 Data at √s = 13 TeV, ATL-PHYS-PUB-2015-051, 2015, URL: https://cds.cern.ch/record/2110140.
  76. [76]J. Alwall et al., The automated computation of tree-level and next-to-leading order differential cross sections, and their matching to parton shower simulations, JHEP 07 (2014) 079, arXiv: 1405.0301 [hep-ph].DOI
  77. [77]R. Frederix and S. Frixione, Merging meets matching in MC@NLO, JHEP 12 (2012) 061, arXiv: 1209.6215 [hep-ph].DOI
  78. [78]ATLAS Collaboration, Modelling and computational improvements to the simulation of single vector-boson plus jet processes for the ATLAS experiment, JHEP 08 (2022) 089, arXiv: 2112.09588 [hep-ex].DOI
  79. [79]T. Sjöstrand et al., An introduction to PYTHIA 8.2, Comput. Phys. Commun. 191 (2015) 159, arXiv: 1410.3012 [hep-ph].DOI
  80. [80]E. Bothmann et al., Event generation with Sherpa 2.2, SciPost Phys. 7 (2019) 034, arXiv: 1905.09127 [hep-ph].DOI
  81. [81]NNPDF Collaboration, R. D. Ball, et al., Parton distributions for the LHC run II, JHEP 04 (2015) 040, arXiv: 1410.8849 [hep-ph].
  82. [82]M. Bähr et al., Herwig++ physics and manual, Eur. Phys. J. C 58 (2008) 639, arXiv: 0803.0883 [hep-ph].DOI
  83. [83]J. Bellm et al., Herwig 7.0/Herwig++ 3.0 release note, Eur. Phys. J. C 76 (2016) 196, arXiv: 1512.01178 [hep-ph].DOI
  84. [84]J. Baglio et al., VBFNLO: A Parton Level Monte Carlo for Processes with Electroweak Bosons – Manual for Version 3.0, (2014), arXiv: 1107.4038 [hep-ph].arxiv.org/abs/1107.4038
  85. [85]L. A. Harland-Lang, A. D. Martin, P. Motylinski, and R. S. Thorne, Parton distributions in the LHC era: MMHT 2014 PDFs, Eur. Phys. J. C 75 (2015) 204, arXiv: 1412.3989 [hep-ph].DOI
  86. [86]S. Alioli, P. Nason, C. Oleari, and E. Re, A general framework for implementing NLO calculations in shower Monte Carlo programs: the POWHEG BOX, JHEP 06 (2010) 043, arXiv: 1002.2581 [hep-ph].DOI
  87. [87]S. Agostinelli et al., GEANT4 – a simulation toolkit, Nucl. Instrum. Meth. A 506 (2003) 250.
  88. [88]ATLAS Collaboration, The ATLAS Simulation Infrastructure, Eur. Phys. J. C 70 (2010) 823, arXiv: 1005.4568 [physics.ins-det].DOI
  89. [89]B. Nachman and J. Thaler, Neural resampler for Monte Carlo reweighting with preserved uncertainties, Phys. Rev. D 102 (2020) 076004, arXiv: 2007.11586 [hep-ph].DOI
  90. [90]A. Ramdas, N. Garcia, and M. Cuturi, On Wasserstein Two Sample Testing and Related Families of Nonparametric Tests, Entropy 19 (2015) 47, arXiv: 1509.02237 [stat.ML].arxiv.org/abs/1509.02237
  91. [91]H. Qu, C. Li, and S. Qian, Particle Transformer for Jet Tagging, (2024), arXiv: 2202.03772 [hep-ph].arxiv.org/abs/2202.03772
  92. [92]V. Mikuni and B. Nachman, Solving key challenges in collider physics with foundation models, Phys. Rev. D 111 (2025) L051504, arXiv: 2404.16091 [hep-ph].arxiv.org/abs/2404.16091
  93. [93]W. Bhimji, C. Harris, V. Mikuni, and B. Nachman, Foundation model framework for all tasks involving jet physics, Phys. Rev. D 113 (3 2026) 032020, URL: https://link.aps.org/doi/10.1103/prd-f5jm.DOI
  94. [94]J. Bardhan, R. Agrawal, A. Tilak, C. Neeraj, and S. Mitra, HEP-JEPA: A foundation model for collider physics using joint embedding predictive architecture, (2025), arXiv: 2502.03933 [cs.LG].arxiv.org/abs/2502.03933
  95. [95]M. Leigh et al., Is tokenization needed for masked particle modeling? Mach. Learn.: Sci. Technol. 6 (2025) 025075, URL: https://doi.org/10.1088/2632-2153/addb98.
  96. [96]T. Golling et al., Masked particle modeling on sets: towards self-supervised high energy physics foundation models, Mach. Learn.: Sci. Technol. 5 (2024) 035074, arXiv: 2401.13537 [hep-ph], URL: https://doi.org/10.1088/2632-2153/ad64a8.
  97. [97]P. Harris, J. Krupa, M. Kagan, B. Maier, and N. Woodward, Resimulation-based self-supervised learning for pretraining physics foundation models, Phys. Rev. D 111 (2025) 032010, arXiv: 2403.07066 [hep-ph].
  98. [98]A. Hallin, “Foundation models for high-energy physics,” 2nd European AI for Fundamental Physics Conference, 2025, arXiv: 2509.21434 [hep-ph].arxiv.org/abs/2509.21434
  99. [99]J. Birk, A. Hallin, and G. Kasieczka, OmniJet-α: the first cross-task foundation model for particle physics, Mach. Learn.: Sci. Technol. 5 (2024) 035031, arXiv: 2403.05618 [hep-ph].DOI
  100. [100]T.-H. Hsu et al., EyeNet: A Foundation Model for Particle Collision Data Analysis, (2026), arXiv: 2601.17126 [hep-ex].DOI
  101. [101]J. Birk et al., Enhancing next token prediction based pre-training for jet foundation models, (2025), arXiv: 2512.04149 [hep-ph].DOI
  102. [102]ATLAS Collaboration, An implementation of neural simulation-based inference for parameter estimation in ATLAS, Rept. Prog. Phys. 88 (2025) 067801, arXiv: 2412.01600 [hep-ex].DOI
  103. [103]ATLAS Collaboration, Measurement of off-shell Higgs boson production in the H* → ZZ → 4ℓ decay channel using a neural simulation-based inference technique in 13 TeV pp collisions with the ATLAS detector, Rept. Prog. Phys. 88 (2025) 057803, arXiv: 2412.01548 [hep-ex].DOI
  104. [104]A. Paszke et al., PyTorch: An Imperative Style, High-Performance Deep Learning Library, Adv. Neural Inf. Process. Syst. 32 (2019) 8024, url: https://proceedings.neurips.cc/paper_files/paper/2019/file/bdbca288fee7f92f2bfa9f7012727740-Paper.pdf.arxiv.org/abs/1912.01703
  105. [105]X. Chen et al., Symbolic Discovery of Optimization Algorithms, Adv. Neural Inf. Process. Syst. 36 (2023) 49205, url: https://proceedings.neurips.cc/paper_files/paper/2023/file/9edb0b98cf4d7c1f3f6f6f8e8c4d8f06-Paper-Conference.pdf.
  106. [106]ATLAS Collaboration, Evaluating statistical uncertainties and correlations using the bootstrap method, ATL-PHYS-PUB-2021-011, 2021, URL: https://cds.cern.ch/record/2759945.
  107. [107]ATLAS Collaboration, Measurement of the Z/γ* boson transverse momentum distribution in pp collisions at √s = 7 TeV with the ATLAS detector, JHEP 09 (2014) 145, arXiv: 1406.3660 [hep-ex].DOI
  108. [108]ATLAS Collaboration, Multi-boson simulation for 13 TeV ATLAS analyses, ATL-PHYS-PUB-2016-002, 2016, URL: https://cds.cern.ch/record/2119986.
  109. [109]ATLAS Collaboration, Measurement of the electroweak production of dijets in association with a Z-boson and distributions sensitive to vector boson fusion in proton–proton collisions at √s = 8 TeV using the ATLAS detector, JHEP 04 (2014) 031, arXiv: 1401.7610 [hep-ex].DOI
  110. [110]ATLAS Collaboration, Study of jet shapes in inclusive jet production in pp collisions at √s = 7 TeV using the ATLAS detector, Phys. Rev. D 83 (2011) 052003, arXiv: 1101.0070 [hep-ex].DOI
  111. [111]CMS Collaboration, Measurement of jet substructure observables in tt̄ events from proton–proton collisions at √s = 13 TeV, Phys. Rev. D 98 (2018) 092014, arXiv: 1808.07340 [hep-ex].DOI
  112. [112]ATLAS Collaboration, Measurement of jet substructure in boosted t\bar{t} events with the ATLAS detector using 140 fb^{-1} of 13 TeV pp collisions, Phys. Rev. D 109 (2024) 112016, arXiv: 2312.03797 [hep-ex].DOI
  113. [113]M. Cacciari and G. P. Salam, Pileup subtraction using jet areas, Phys. Lett. B 659 (2008) 119, arXiv: 0707.1378 [hep-ph].DOI
  114. [114]M. Cacciari, G. P. Salam, and G. Soyez, The Catchment Area of Jets, JHEP 04 (2008) 005, arXiv: 0802.1188 [hep-ph].DOI
  115. [115]M. Cacciari, G. P. Salam, and G. Soyez, FastJet user manual, Eur. Phys. J. C 72 (2012) 1896, arXiv: 1111.6097 [hep-ph].DOI
  116. [116]CMS Collaboration, Study of quark and gluon jet substructure in Z+jet and dijet events from pp collisions, JHEP 2022 (2022) 188, URL: https://doi.org/10.1007/JHEP01(2022)188.
  117. [117]I. Moult and H. Xing Zhu, Energy Correlators: A Journey From Theory to Experiment, (2025), arXiv: 2506.09119 [hep-ph].arxiv.org/abs/2506.09119
  118. [118]M. Jaarsma, Y. Li, I. Moult, W. J. Waalewijn, and H. Xing Zhu, From DGLAP to Sudakov: Precision Predictions for Energy-Energy Correlators, (2025), arXiv: 2512.11950 [hep-ph].arxiv.org/abs/2512.11950
  119. [119]W. Chen et al., NNLL resummation for projected three-point energy correlator, JHEP 05 (2024) 043, arXiv: 2307.07510 [hep-ph].arxiv.org/abs/2307.07510
  120. [120]CMS Collaboration, Measurement of Energy Correlators inside Jets and Determination of the Strong Coupling α_s(m_Z), Phys. Rev. Lett. 133 (2024) 071903, arXiv: 2402.13864 [hep-ex].DOI
  121. [121]K. Lee, B. Meçaj, and I. Moult, Conformal collider physics meets LHC data, Phys. Rev. D 111 (2025) L011502, arXiv: 2205.03414 [hep-ph].
  122. [122]SLD Collaboration, Measurement of α_s(M_Z^2) from hadronic event observables at the Z^0 resonance, Phys. Rev. D 51 (1995) 962, arXiv: hep-ex/9501003.
  123. [123]DELPHI Collaboration, Energy-energy correlations in hadronic final states from Z0 decays, Phys. Lett. B 252 (1990) 149.
  124. [124]L3 Collaboration, Determination of α_s from energy-energy correlations measured on the Z0 resonance., Phys. Lett. B 257 (1991) 469.
  125. [125]ALEPH Collaboration, Measurement of α_s from the structure of particle clusters produced in hadronic Z decays, Phys. Lett. B 257 (1991) 479.
  126. [126]OPAL Collaboration, A Measurement of energy correlations and a determination of α_s(M_Z^2) in e+ e- annihilations at √s = 91 GeV, Phys. Lett. B 252 (1990) 159.
  127. [127]ATLAS Collaboration, Measurement of transverse energy–energy correlations in multi-jet events in pp collisions at √s = 7 TeV using the ATLAS detector and determination of the strong coupling constant α_s(m_Z), Phys. Lett. B 750 (2015) 427, arXiv: 1508.01579 [hep-ex].DOI
  128. [128]ATLAS Collaboration, Determination of the strong coupling constant αs from transverse energy-energy correlations in multijet events at √s = 8 TeV using the ATLAS detector, Eur. Phys. J. C 77 (2017) 872, arXiv: 1707.02562 [hep-ex].DOI
  129. [129]ATLAS Collaboration, Determination of the strong coupling constant from transverse energy–energy correlations in multijet events at √s = 13 TeV with the ATLAS detector, JHEP 07 (2023) 085, arXiv: 2301.09351 [hep-ex].DOI
  130. [130]P. T. Komiske, I. Moul, J. Thaler, and H. X. Zhu, Analyzing N-Point Energy Correlators inside Jets with CMS Open Data, Phys. Rev. Lett. 130 (2023) 051901, arXiv: 2201.07800 [hep-ph].DOI
  131. [131]A. Badea et al., Measurements of two-particle correlations in e+e− collisions at 91 GeV with ALEPH archived data, Phys. Rev. Lett. 123 (2019) 212002, arXiv: 1906.00489 [hep-ex].DOI
  132. [132]H. Bossi et al., Energy Correlators from Partons to Hadrons: Unveiling the Dynamics of the Strong Interactions with Archival ALEPH Data, (2025), arXiv: 2511.00149 [hep-ph].arxiv.org/abs/2511.00149
  133. [133]STAR Collaboration, Measurement of Two-Point Energy Correlators within Jets in p+p Collisions at √s = 200 GeV, Phys. Rev. Lett. 135 (2025) 111901, arXiv: 2502.15925 [hep-ex].arxiv.org/abs/2502.15925
  134. [134]CMS Collaboration, Energy-energy correlators from PbPb and pp collisions at 5.02 TeV, CMS Public Analysis Summary CMS-PAS-HIN-23-004, CERN, 2024, URL: https://cds.cern.ch/record/2905611.
  135. [135]ALICE Collaboration, Exposing the parton-hadron transition within jets with energy-energy correlators in pp collisions at √s = 5.02 TeV, (2024), arXiv: 2409.12687 [hep-ex].DOI
  136. [136]F. Camastra and A. Staiano, Intrinsic Dimension Estimation: Advances and Open Problems, Information Sciences 328 (2016) 26, ISSN: 0020-0255.DOI
  137. [137]P. T. Komiske, E. M. Metodiev, and J. Thaler, Metric Space of Collider Events, Phys. Rev. Lett. 123 (2019) 041801, arXiv: 1902.02346 [hep-ph].DOI
  138. [138]R. Tito D’Agnolo, A. Glioti, G. Rigo, and A. Valenti, The Intrinsic Dimension of Collider Events and Model-Independent Searches in 100 Dimensions, (2025), arXiv: 2511.20760 [hep-ph].arxiv.org/abs/2511.20760
  139. [139]E. Levina and P. Bickel, “Maximum Likelihood Estimation of Intrinsic Dimension,” Advances in Neural Information Processing Systems, ed. by L. Saul, Y. Weiss, and L. Bottou, vol. 17, MIT Press, 2004, URL: https://proceedings.neurips.cc/paper_files/paper/2004/file/74934548253bcab8490ebd74afed7031-Paper.pdf.
  140. [140]M. Cacciari, G. P. Salam, and G. Soyez, The anti-kt jet clustering algorithm, JHEP 04 (2008) 063, arXiv: 0802.1189 [hep-ph].DOI
  141. [141]ATLAS Collaboration, ATLAS Computing Acknowledgements, ATL-SOFT-PUB-2026-001, 2026, URL: https://cds.cern.ch/record/2952666.
  142. [142]J. Thaler and K. Van Tilburg, Identifying Boosted Objects with N-subjettiness, JHEP 03 (2011) 015, arXiv: 1011.2268 [hep-ph].DOI

Paper details

vlnc:W_cw6cn9cv
Version
v1 · latest
Published
Fresh — added to the Hub within the last 30 days
Added
License
CC BY 4.0
Usage

views · 2

downloads · 0

A view counts once per browser per day, a download each time the PDF is requested.

  • Physics and Astronomy

Not recorded for this paper

View statement
Looking for related papers…
Contents