The identification of boosted heavy particles such as top quarks or vector bosons is one of the key problems arising in experimental studies at the Large Hadron Collider. In this article, we introduce LundNet, a novel jet tagging method which relies on graph neural networks and an efficient description of the radiation patterns within a jet to optimally disentangle signatures of boosted objects from background events. We apply this framework to a number of different benchmarks, showing significantly improved performance for top tagging compared to existing state-of-the-art algorithms. We study the robustness of the LundNet taggers to non-perturbative and detector effects, and show how kinematic cuts in the Lund plane can mitigate overfitting of the neural network to model-dependent contributions. Finally, we consider the computational complexity of this method and its scaling as a function of kinematic Lund plane cuts, showing an order of magnitude improvement in speed over previous graph-based taggers.
Lawrence Berkeley National Laboratory（劳伦斯伯克利国家实验室）；Princeton University（普林斯顿大学）；University of California Berkeley（加州大学伯克利分校）
Composing Graph Neural Networks (GNNs) of operations that respect physical symmetries has been suggested to give better model performance with a smaller number of learnable parameters. However, real-world applications, such as in high energy physics have not born this out. We present the novel architecture VecNet that combines both symmetry-respecting and unconstrained operations to study and tune the degree of physics-informed GNNs. We introduce a novel metric, the \textit{ant factor}, to quantify the resource-efficiency of each configuration in the search-space. We find that a generalized ar…
We study the boosted Higgs tagging using the Lund jet plane. The convolutional neural network is used for the Lund images data set to classify hadronically decaying Higgs from the QCD background. We consider $H\to b \bar{b}$ and $H \to gg$ decay for moderate and high Higgs transverse momentum and compare the performance with the cut based approach using the jet color ring observable. The approach using Lund plane images provides good tagging efficiency for all the cases.
We study the problem of distinguishing $b$-jets stemming from the decay of a colour singlet, such as the Higgs boson, from those originating from the abundant QCD background. In particular, as a case study, we focus on associate production of a vector boson and a Higgs boson decaying into a pair of $b$-jets, which has been recently observed at the LHC. We consider the combination of several theory-driven observables proposed in the literature, together with Lund jet plane images, in order to design an original $Hbb$ tagger. The observables are combined by means of standard machine learning alg…
We construct a procedure to separate boosted Higgs bosons decaying into hadrons, from the background due to strong interactions. We employ the Lund jet plane to obtain a theoretically well-motivated representation of the jets of interest and we use the resulting images as the input to a convolutional neural network classifier. In particular, we consider two different decay modes of the Higgs boson, namely into a pair of bottom quarks or into light jets, against the respective backgrounds. For each case, we consider both a moderate- and high- boost scenario. The performance of the tagger is com…
Lund diagrams, a theoretical representation of the phase space within jets, have long been used in discussing parton showers and resummations. We point out that they can be created for individual jets through repeated Cambridge/Aachen declustering, providing a powerful visual representation of the radiation within any given jet. Concentrating here on the primary Lund plane, we outline some of its analytical properties, highlight its scope for constraining Monte Carlo simulations and comment on its relation with existing observables such as the $z_g$ variable and the iterated soft-drop multipli…
The Lund-jet plane has recently been proposed as a powerful jet substructure tool with a broad range of applications. In this paper, we provide an all-order single logarithmic calculation of the primary Lund-plane density in Quantum Chromodynamics, including contributions from the running of the coupling, collinear effects for the leading parton, and soft logarithms that account for large-angle and clustering effects. We also identify a new source of clustering logarithms close to the boundary of the jet, deferring their resummation to future work. We then match our all-order results to exact …
Technical University of Munich, TUM School of Natural Sciences, Physics Department（慕尼黑工业大学）；Dipartimento di Fisica, Università di Genova and INFN, Sezione di Genova（热那亚大学物理系和意大利国家核物理研究所热那亚分部）；Institut de Physique Théorique, Paris Saclay University, CNRS, CEA（巴黎萨克雷大学理论物理研究所）
Università degli Studi di Milano（米兰大学）；INFN（意大利国家核物理研究所）；University of Oxford（牛津大学）
We introduce a generative model to simulate radiation patterns within a jet using the Lund jet plane. We show that using an appropriate neural network architecture with a stochastic generation of images, it is possible to construct a generative model which retrieves the underlying two-dimensional distribution to within a few percent. We compare our model with several alternative state-of-the-art generative techniques. Finally, we show how a mapping can be created between different categories of jets, and use this method to retroactively change simulation settings or the underlying process on a…
We study the issue of separating hadronic jets that contain bottom quarks ($b$-jets) from jets featuring light partons only. We develop a novel approach to $b$-tagging that exploits the application of QCD-inspired jet substructure observables such as one-dimensional jet angularities and the two-dimensional primary Lund plane. We demonstrate that these observables can be used as inputs to modern machine-learning algorithms to efficiently separate $b$-jets from light ones. In order to test our tagging procedure, we consider simulated events where a $Z$ boson is produced is association with jets …
University of California, Santa Barbara（加州大学圣巴巴拉分校）；CERN（欧洲核子研究组织）
How to represent a jet is at the core of machine learning on jet physics. Inspired by the notion of point clouds, we propose a new approach that considers a jet as an unordered set of its constituent particles, effectively a "particle cloud". Such a particle cloud representation of jets is efficient in incorporating raw information of jets and also explicitly respects the permutation symmetry. Based on the particle cloud representation, we propose ParticleNet, a customized neural network architecture using Dynamic Graph Convolutional Neural Network for jet tagging problems. The ParticleNet arc…
We present a new technique for tagging heavy-flavor jets with p_T > 500 GeV called "mu_x tagging." Current track-based methods of b-jet tagging lose efficiency and experience a large rise in fake rate in the boosted regime. Using muons from B hadron decay, we combine angular information and jet substructure to tag b jets, c jets, light jets, and "light-heavy" jets (those containing B hadrons from gluon splitting). We find tagging efficiencies of epsilon_b = 14%, epsilon_c = 6.5%, epsilon_{light-light} = 0.14%, and epsilon_{light-heavy} = 0.5%, respectively, that are nearly independent of trans…
We revisit the calculation of the average jet multiplicity in high-energy collisions. First, we introduce a new definition of (sub)jet multiplicity based on Lund declusterings obtained using the Cambridge jet algorithm. We develop a new systematic resummation approach. This allows us to compute both the Lund and the Cambridge average multiplicities to next-to-next-to-double (NNDL) logarithmic accuracy in electron-positron annihilation, an order higher in accuracy than previous works in the literature. We match our resummed calculation to the exact NLO ($\mathcal{O}(α_s^2)$) result, showing pre…
Mila — Quebec Artificial Intelligence Institute（米拉——魁北克人工智能研究所）；Université de Montréal（蒙特利尔大学）
We leverage representation learning and the inductive bias in neural-net-based Standard Model jet classification tasks, to detect non-QCD signal jets. In establishing the framework for classification-based anomaly detection in jet physics, we demonstrate that, with a \emph{well-calibrated} and \emph{powerful enough feature extractor}, a well-trained \emph{mass-decorrelated} supervised Standard Model neural jet classifier can serve as a strong generic anti-QCD jet tagger for effectively reducing the QCD background. Imposing \emph{data-augmented} mass-invariance (and thus decoupling the dominant…
I present a new scheme for tagging boosted heavy flavor jets called "$μ_x$ tagging" and its application to TeV-scale physics beyond the Standard Model. Using muons from B hadron decay to define a particular combination "x" of angular information, and jet substructure variables, we identify a clean ($ε_{\mathrm{fake}}/ε_b\sim 1/100$) good efficiency ($ε_b = 14\%$) tag. I demonstrate the usefulness of this new scheme by showing the reach for discovery of leptophobic $Z^\prime\to b\bar b$ and $tH^\pm\to t t b$.
We present a new scheme for tagging high-$p_{T}$ bottom and charm jets using energetic muons. Contemporary track-based $b$ tags lose their ability to reject light jet background as jet $p_{T}\rightarrow\mathcal{O}(\mathrm{TeV})$, where the massive boost exposes fundamental limits in tracking resolution. For our "$μ_{x}$" tag, the signal efficiency and light jet rejection is robust versus $p_{T}$. In the tested regime (jet $p_{T}\in[\mathrm{0.5,2.1}]$ TeV), $μ_{x}$ tags $\sim14\%$ of $b$ jets, $\sim6.5\%$ of $c$ jets and $\sim0.65\%$ of light jets. Since $μ_{x}$ tagging should be immediately us…
Recent literature on deep neural networks for tagging of highly energetic jets resulting from top quark decays has focused on image based techniques or multivariate approaches using high-level jet substructure variables. Here, a sequential approach to this task is taken by using an ordered sequence of jet constituents as training inputs. Unlike the majority of previous approaches, this strategy does not result in a loss of information during pixelisation or the calculation of high level features. The jet classification method achieves a background rejection of 45 at a 50% efficiency operating …
Convolutional neural networks are basic structures using jet images as input for the jet tagging problems. However, what they have learned during the training process is always difficult to understand just through feature maps. Inspired by the attention mechanism popular in machine learning fields, we propose a novel attention-based neural network (ABNN) to get insight of this problem. The ABNN combines a jet image with average jet images from the signal and the background to generate attention maps which show clearly the relevant importance according to the different origination of jets. Comp…
Heavy flavour jet tagging is widely used in the determination of cross sections including the production of heavy flavoured quarks. This requires the knowledge of heavy and light flavour jet tagging efficiencies and their uncertainties. A system of eight non-linear equations can be used to determine these quantities by means of two different tagging algorithms and two data sets which differ in heavy flavour jet content. The analytical solution of this system of equations, derived by means of resultants is described in detail, including a discussion about its singularities which provide importa…
Deep learning methods have been increasingly adopted to study jets in particle physics. Since symmetry-preserving behavior has been shown to be an important factor for improving the performance of deep learning in many applications, Lorentz group equivariance - a fundamental spacetime symmetry for elementary particles - has recently been incorporated into a deep learning model for jet tagging. However, the design is computationally costly due to the analytic construction of high-order tensors. In this article, we introduce LorentzNet, a new symmetry-preserving deep learning model for jet taggi…
We propose a method to probe the coupling of the Higgs to strange quarks by tagging strange jets at future lepton colliders. For this purpose we describe a jet-flavor observable, $J_F$, that is correlated with the flavor of the quark associated with the hard part of the jet. Using this variable, we set up a strangeness tagger aimed at studying the decay $h\to s\bar{s}$. We determine the sensitivity of our method to the strange Yukawa coupling, and find it to be of the order of the standard-model expectation.
We introduce a novel approach to jet tagging and classification through the use of techniques inspired by computer vision. Drawing parallels to the problem of facial recognition in images, we define a jet-image using calorimeter towers as the elements of the image and establish jet-image preprocessing methods. For the jet-image processing step, we develop a discriminant for classifying the jet-images derived using Fisher discriminant analysis. The effectiveness of the technique is shown within the context of identifying boosted hadronic W boson decays with respect to a background of quark- and…
We consider the process $p p \to t {\bar t} H$. This process can give rise to many signatures of the Higgs boson. The signatures can have electrons, muons and jets. We consider the signatures that have two electrons/muons and jets. Tagging of a tau jet and a bottom jet can help reduce the backgrounds significantly. In particular, we examine the usefulness of the signatures "isolated 2 electrons/muons + a bottom jet + a tau jet", "isolated 2 electrons/muons + 2 tau jets", "isolated 2 electrons/muons + 2 bottom jets + a tau jet", and "isolated 2 electrons/muons + a bottom jet + 2 tau jets". We f…
We introduce a novel anomaly search method based on (i) jet tagging to select interesting events, which are less likely to be produced by background processes; (ii) comparison of the untagged and tagged samples to single out features (such as bumps produced by the decay of new particles) in the latter. We demonstrate the usefulness of this method by applying it to a final state with two massive boosted jets: for the new physics benchmarks considered, the signal significance increases an order of magnitude, up to a factor of 40. We compare to other anomaly detection methods in the literature an…
Jet tagging has become an essential tool for new physics searches at the high-energy frontier. For jets that contain energetic charged leptons we introduce Feature Extended Supervised Tagging (FEST) which, in addition to jet substructure, considers the features of the charged lepton within the jet. With this method we build dedicated taggers to discriminate among boosted $H \to \ell νq \bar q$, $t \to \ell νb$, and QCD jets (with $\ell$ an electron or muon). The taggers have an impressive performance, allowing for overall light jet rejection factors of $10^4-10^5$, for top quark / Higgs boson …
Jet flavour identification algorithms are of paramount importance to maximise the physics potential of future collider experiments. This work describes a novel set of tools allowing for a realistic simulation and reconstruction of particle level observables that are necessary ingredients to jet flavour identification. An algorithm for reconstructing the track parameters and covariance matrix of charged particles for an arbitrary tracking sub-detector geometries has been developed. Additional modules allowing for particle identification using time-of-flight and ionizing energy loss information …
We present a new tagger which aims at identifying partially reconstructed objects, in which only some of the constituents are collected in a single jet. As an example, we focus on top decays in which either part of the hadronically decaying W or the b jet is soft or falls outside of the top jet cone. We construct an observable to identify remnant substructure from the decay and employ aggressive jet grooming to reject QCD backgrounds. The tagger is complementary to existing ones and works well in the intermediate boost regime where jet substructure techniques usually fail. It is anticipated th…
Tagging jets of strongly interacting particles initiated by energetic strange quarks is one of the few largely unexplored Standard Model object classification problems remaining in high energy collider physics. In this paper we investigate the purest version of this classification problem in the form of distinguishing strange-quark jets from down-quark jets. Our strategy relies on the fact that a strange-quark jet contains on average a higher ratio of neutral kaon energy to neutral pion energy than does a down-quark jet. Long-lived neutral kaons deposit energy mainly in the hadronic calorimete…
We initiate the study of the time substructure of jets, motivated by the fact that the next generation of detectors at particle colliders will resolve the time scale over which jet constituents arrive. This effect is directly related to hadronization, which transforms partons into massive hadrons with a distribution of velocities. We review the basic predictions for the velocity distribution of jet hadrons, and suggest an application for this information in the context of boosted object tagging. By noting that the velocity distribution is determined by the properties of the color string which …
A method is proposed for distinguishing highly boosted hadronically decaying W's (W-jets) from QCD-jets using jet substructure. Previous methods, such as the filtering/mass-drop method, can give a factor of ~2 improvement in S/sqrt(B) for jet pT > 200 GeV. In contrast, a multivariate approach including new discriminants such as R-cores, which characterize the shape of the W-jet, subjet planar flow, and grooming-sensitivities is shown to provide a much larger factor of ~5 improvement in S/sqrt(B). For longitudinally polarized W's, such as those coming from many new physics models, the discrimin…
The separation of quark and gluon initiated jets can be an important way to improve the sensitivity in searches for new physics or in measurements of Higgs boson properties. We present a simplified version of the shower deconstruction approach as a novel observable for quark-gluon tagging. Assuming topocluster-like objects as input, we compare our observable with energy correlation functions and find a favorable performance for a large variety of jet definitions. When applied to dark matter searches in mono-jet final states limitations from small signal-to-background ratios can be overcome. We…
Searches for invisible Higgs decays in weak boson fusion are a well-known laboratory for jets and QCD studies. We present a series of results on tagging jets and central jet activity. First, precision analyses of the central jet activity require full control of single top production in some analyses. Second, the rate dependence on the size of the tagging jets is not limited to weak boson fusion. For the first time, we show how subjet information on the tagging jets and on the additional jet activity can be used to extract the Higgs signal. The additional observables relieve some of the pressur…
A new jet observable, dipolarity, is introduced that can distinguish whether a pair of subjets arises from a color singlet source. This observable is incorporated into the HEPTopTagger and is shown to improve discrimination between top jets and QCD jets for moderate to high pT.
The main parts of the LHC diffractive physics programme possible to be measured using a proton tagging technique are presented. The geometric acceptance of the ATLAS forward proton detectors: ALFA and AFP for various LHC optics settings are shown. The probabilities of observing a proton originating from a minimum-bias event in ALFA and AFP stations are given. The main properties of single diffractive and double Pomeron exchange production of dijets, photon+jet, jet-gap-jet and W/Z bosons are discussed. The possibility of measuring the jet production in exclusive (double proton tag) and semi-ex…
Deep learning is a standard tool in the field of high-energy physics, facilitating considerable sensitivity enhancements for numerous analysis strategies. In particular, in identification of physics objects, such as jet flavor tagging, complex neural network architectures play a major role. However, these methods are reliant on accurate simulations. Mismodeling can lead to non-negligible differences in performance in data that need to be measured and calibrated against. We investigate the classifier response to input data with injected mismodelings and probe the vulnerability of flavor tagging…
Jet identification tools are crucial for new physics searches at the LHC and at future colliders. We introduce the concept of Mass Unspecific Supervised Tagging (MUST) which relies on considering both jet mass and transverse momentum varying over wide ranges as input variables - together with jet substructure observables - of a multivariate tool. This approach not only provides a single efficient tagger for arbitrary ranges of jet mass and transverse momentum, but also an optimal solution for the mass correlation problem inherent to current taggers. By training neural networks, we build MUST-i…
$N$-subjettiness ratios are in wide use for tagging heavy boosted objects, in particular the ratio of 2-subjettiness to 1-subjettiness for tagging boosted electroweak bosons. In this article we introduce a new, \emph{dichroic} ratio, which uses different regions of a jet to determine the two subjettiness measures, emphasising the hard substructure for the 1-subjettiness and the full colour radiation pattern for the 2-subjettiness. Relative to existing $N$-subjettiness ratios, the dichroic extension, combined with SoftDrop (pre-)grooming, makes it possible to increase the ultimate signal signif…
Jet flavor tagging with Particle Transformer for Higgs factories
International Center for Elementary Particle Physics, The University of Tokyo（东京大学基础粒子物理国际中心）；Department of Physics, Graduate School of Science, The University of Tokyo（东京大学理学研究科物理学系）；MI-6 Ltd.（MI-6有限公司）
We study the production of an electroweak boson in association with jets, in processes where the jet with the highest transverse momentum is identified as quark-initiated. The quark/gluon tagging procedure is realised by a cut on a jet angularity and it is therefore theoretically well-defined and exhibits infrared and collinear safety. In this context, exploiting resummed perturbation theory, we are able to provide theoretical predictions for transverse momentum distributions at a well-defined and, in principle, systematically improvable accuracy. In particular, tagging the leading jet as quar…
The University of British Columbia（不列颠哥伦比亚大学）；McGill University（麦吉尔大学）；Stanford University（斯坦福大学）
Multivariate techniques based on engineered features have found wide adoption in the identification of jets resulting from hadronic top decays at the Large Hadron Collider (LHC). Recent Deep Learning developments in this area include the treatment of the calorimeter activation as an image or supplying a list of jet constituent momenta to a fully connected network. This latter approach lends itself well to the use of Recurrent Neural Networks. In this work the applicability of architectures incorporating Long Short-Term Memory (LSTM) networks is explored. Several network architectures, methods …
We propose a novel method in that quark-gluon tagging of the jets emitted as initial state radiation (ISR) can boost searches of invisible Higgs from gluon fusion processes against irreducible electroweak vector boson productions. While quark ISR typically takes up a dominant portion than gluon in the background processes mainly by frequent quark-gluon initiated hard scatterings at the LHC, gluon ISR portion in the gluon fusion can be significantly larger in the central region of detector. Focusing on invisible Higgs searches using jet substructure variables capturing the new features, we demo…
The signature produced by the Standard Model Higgs boson in the Vector Boson Fusion (VBF) mechanism is usually pinpointed by requiring two well separated hadronic jets, one of which (at least) of them tends to be in the forward direction. With the increase of instantaneous luminosity at the LHC, the isolation of the Higgs boson produced with the VBF mechanism is rendered more challenging. In this paper the feasibility of single jet tagging is explored in a high-luminosity scenario. It is demonstrated that the separation in rapidity between the tagging jet and the Higgs boson can be effectively…
We introduce a method to discover the Higgs boson at the Large Hadron Collider (LHC) through its decay to W^+W^-, where one boson decays to leptons, and the other decays to c+jet. This mechanism is complementary to the decay into dileptons, but has the potential to measure the invariant mass peak of the Higgs boson, and to avoid large recently discovered QCD backgrounds from heavy flavor decays. In addition, this mechanism motivates the study and creation of a dedicated charm jet tagger at LHC experiments. Existing charm jet tagging, in the form of fakes to bottom jet tagging, provides sensiti…
