arXiv subjects

Kyoungchul Kong

Publications and source records attributed to Kyoungchul Kong.

At least 37 records · Page 2Linked to original sources

arXiv2024

When the Machine Chimes the Bell: Entanglement and Bell Inequalities with Boosted $t\bar{t}$

The Large Hadron Collider provides a unique opportunity to study quantum entanglement and violation of Bell inequalities at the highest energy available today. In this paper, we will investigate these quantum correlations with top quark pair production, which represents a system of two-qubits. The spacelike separation requirement for the two causally disconnected top quarks requires they fly relativistically away from each other, which motivates the use of the boosted top-tagging with the semi-leptonic top pair channel. Although measuring the spin polarization of the hadronic top quark is known to be challenging, our study indicates that it is feasible to reconstruct the spin density matrix of the two-qubit system using an optimal hadronic polarimeter. This is achieved with the aid of jet substructure techniques and NN-inspired reconstruction methods, which improve the mapping between subjets and quarks. We find that entanglement can already be observed at more than $5σ$ level with existing data, and violation of Bell inequalities may be probed above 4$σ$ level at the HL-LHC with 3 ab$^{-1}$ of data.

hep-ph↗

arXiv2024

Hadronic Top Quark Polarimetry with ParticleNet

Precision studies for top quark physics are a cornerstone of the Large Hadron Collider program. Polarization, probed through decay kinematics, provides a unique tool to scrutinize the top quark across its various production modes and to explore potential new physics effects. However, the top quark most often decays hadronically, for which unambiguous identification of its decay products sensitive to top quark polarization is not possible. In this Letter, we introduce a jet flavor tagging method to significantly improve spin analyzing power in hadronic decays, going beyond exclusive kinematic information employed in previous studies. We provide parametric estimates of the improvement from flavor tagging with any set of measured observables and demonstrate this in practice on simulated data using a Graph Neural Network (GNN). We find that the spin analyzing power in hadronic decays can improve by approximately 20% (40%) compared to the kinematic approach, assuming an efficiency of 0.5 (0.2) for the network.

hep-ph↗

arXiv2024

A Comparison Between Invariant and Equivariant Classical and Quantum Graph Neural Networks

Machine learning algorithms are heavily relied on to understand the vast amounts of data from high-energy particle collisions at the CERN Large Hadron Collider (LHC). The data from such collision events can naturally be represented with graph structures. Therefore, deep geometric methods, such as graph neural networks (GNNs), have been leveraged for various data analysis tasks in high-energy physics. One typical task is jet tagging, where jets are viewed as point clouds with distinct features and edge connections between their constituent particles. The increasing size and complexity of the LHC particle datasets, as well as the computational models used for their analysis, greatly motivate the development of alternative fast and efficient computational paradigms such as quantum computation. In addition, to enhance the validity and robustness of deep networks, one can leverage the fundamental symmetries present in the data through the use of invariant inputs and equivariant layers. In this paper, we perform a fair and comprehensive comparison between classical graph neural networks (GNNs) and equivariant graph neural networks (EGNNs) and their quantum counterparts: quantum graph neural networks (QGNNs) and equivariant quantum graph neural networks (EQGNN). The four architectures were benchmarked on a binary classification task to classify the parton-level particle initiating the jet. Based on their AUC scores, the quantum networks were shown to outperform the classical networks. However, seeing the computational advantage of the quantum networks in practice may have to wait for the further development of quantum technology and its associated APIs.

quant-ph↗

arXiv2024

Quantum Vision Transformers for Quark-Gluon Classification

We introduce a hybrid quantum-classical vision transformer architecture, notable for its integration of variational quantum circuits within both the attention mechanism and the multi-layer perceptrons. The research addresses the critical challenge of computational efficiency and resource constraints in analyzing data from the upcoming High Luminosity Large Hadron Collider, presenting the architecture as a potential solution. In particular, we evaluate our method by applying the model to multi-detector jet images from CMS Open Data. The goal is to distinguish quark-initiated from gluon-initiated jets. We successfully train the quantum model and evaluate it via numerical simulations. Using this approach, we achieve classification performance almost on par with the one obtained with the completely classical architecture, considering a similar number of parameters.

quant-ph↗

arXiv2024

$\mathbb{Z}_2\times \mathbb{Z}_2$ Equivariant Quantum Neural Networks: Benchmarking against Classical Neural Networks

This paper presents a comprehensive comparative analysis of the performance of Equivariant Quantum Neural Networks (EQNN) and Quantum Neural Networks (QNN), juxtaposed against their classical counterparts: Equivariant Neural Networks (ENN) and Deep Neural Networks (DNN). We evaluate the performance of each network with two toy examples for a binary classification task, focusing on model complexity (measured by the number of parameters) and the size of the training data set. Our results show that the $\mathbb{Z}_2\times \mathbb{Z}_2$ EQNN and the QNN provide superior performance for smaller parameter sets and modest training data samples.

quant-ph↗

arXiv2024

Hybrid Quantum Vision Transformers for Event Classification in High Energy Physics

Models based on vision transformer architectures are considered state-of-the-art when it comes to image classification tasks. However, they require extensive computational resources both for training and deployment. The problem is exacerbated as the amount and complexity of the data increases. Quantum-based vision transformer models could potentially alleviate this issue by reducing the training and operating time while maintaining the same predictive power. Although current quantum computers are not yet able to perform high-dimensional tasks yet, they do offer one of the most efficient solutions for the future. In this work, we construct several variations of a quantum hybrid vision transformer for a classification problem in high energy physics (distinguishing photons and electrons in the electromagnetic calorimeter). We test them against classical vision transformer architectures. Our findings indicate that the hybrid models can achieve comparable performance to their classical analogues with a similar number of parameters.

quant-ph↗

arXiv2024

$M_{TN}$ is all you need: production of multiple semi-invisible resonances at hadron colliders

The stransverse mass variable $M_{T2}$ was originally proposed for the study of hadron collider events in which $N=2$ parent particles are produced and then decay semi-invisibly. Here we consider the generalization to the case of $N\ge 3$ semi-invisibly decaying parent particles. We introduce the corresponding class of kinematic variables $M_{TN}$ and illustrate their mathematical properties. Many of the celebrated features of the $M_{T2}$ kinematic endpoint are retained in this more general case, including the ability to measure the mass of the invisible daughter particle from the stransverse mass kink. We describe and validate a numerical procedure for computing $M_{TN}$ in practice. We also identify the configurations of visible momenta which result in nontrivial ($M_{TN}\ne 0$) values, and derive a pure phase-space estimate for the fraction of such events for any $N$.

hep-ph↗

arXiv2024

Variance Reduction via Simultaneous Importance Sampling and Control Variates Techniques Using Vegas

Monte Carlo (MC) integration is an important calculational technique in the physical sciences. Practical considerations require that the calculations are performed as accurately as possible for a given set of computational resources. To improve the accuracy of MC integration, a number of useful variance reduction algorithms have been developed, including importance sampling and control variates. In this work, we demonstrate how these two methods can be applied simultaneously, thus combining their benefits. We provide a python wrapper, named CoVVVR, which implements our approach in the Vegas program. The improvements are quantified with several benchmark examples from the literature.

hep-ph↗

arXiv2023

Warm Surprises from Cold Duets: N-Body Simulations with Two-Component Dark Matter

We explore extensive N-body simulations with two-component cold dark matter candidates. We delve into the temperature evolution, power spectrum, density perturbation, and maximum circular velocity functions. We find that the substantial mass difference between the two candidates and the annihilation of the heavier components to the lighter ones effectively endow the latter with warm dark matter-like behavior, taking advantage of all distinct features that warm dark matter candidates offer, without observational bounds on the warm dark matter mass. Moreover, we demonstrate that the two-component dark matter model aligns well with observational data, providing valuable insights into where and how to search for the elusive dark matter candidates in terrestrial experiments.

hep-ph↗

arXiv2023

Exploring the Synergy of Kinematics and Dynamics for Collider Physics

In collider experiments, an event is characterized by two distinct yet mutually complementary features: the `global features' and the `local features'. Kinematic information such as the event topology of a hard process, masses, and spins of particles comprises global features spanning the entire phase space. This global feature can be inferred from reconstructed objects. In contrast, representations of particles in gauge groups, such as Quantum Chromodynamics (QCD), offer localized features revealing the dynamics of an underlying theory. These local features, particularly observed in the patterns of radiation as raw data in various detector components, complement the global kinematic features. In this letter, we propose a simple but effective neural network architecture that seamlessly integrates information from both kinematics and QCD to enhance the signal sensitivity at colliders.

hep-ph↗

arXiv2023

Quantum Metric Learning for New Physics Searches at the LHC

In the NISQ (Noisy intermediate-scale quantum) area, Quantum computers can be utilized for deep learning by treating variational quantum circuits as neural network models. This can be achieved by first encoding the input data onto quantum computers using nonparametric unitary gates. An alternative approach is to train the data encoding to map input data from different classes to separated locations in the Hilbert space. The separation is achieved with metric loss functions, hence the naming ``Quantum Metric Learning". With the limited number of qubits in the NISQ area, this approach works naturally as a hybrid classical-quantum computation enabling embedding of high-dimensional feature data into a small number of qubits. Here, we consider an example of the global QCD color structure of hard b-jets emerging from color singlet scalar decays to optimize the signal to background discrimination with a hybrid classical-quantum metric learning. Due to the sparsity of data, self-supervised methods with data augmentation have been utilized so far. Compared to the this classical self-supervised approach, our hybrid method shows the better classification performance without data augmentations. We emphasize that performance enhancements independent of data augmentation techniques are devoid of the artificial risks introduced by data augmentation.

hep-ph↗

arXiv2023

New Machine Learning Techniques for Simulation-Based Inference: InferoStatic Nets, Kernel Score Estimation, and Kernel Likelihood Ratio Estimation

We propose an intuitive, machine-learning approach to multiparameter inference, dubbed the InferoStatic Networks (ISN) method, to model the score and likelihood ratio estimators in cases when the probability density can be sampled but not computed directly. The ISN uses a backend neural network that models a scalar function called the inferostatic potential $φ$. In addition, we introduce new strategies, respectively called Kernel Score Estimation (KSE) and Kernel Likelihood Ratio Estimation (KLRE), to learn the score and the likelihood ratio functions from simulated data. We illustrate the new techniques with some toy examples and compare to existing approaches in the literature. We mention en passant some new loss functions that optimally incorporate latent information from simulations into the training procedure.

stat.ML↗

arXiv2022

Is the Machine Smarter than the Theorist: Deriving Formulas for Particle Kinematics with Symbolic Regression

We demonstrate the use of symbolic regression in deriving analytical formulas, which are needed at various stages of a typical experimental analysis in collider phenomenology. As a first application, we consider kinematic variables like the stransverse mass, $M_{T2}$, which are defined algorithmically through an optimization procedure and not in terms of an analytical formula. We then train a symbolic regression and obtain the correct analytical expressions for all known special cases of $M_{T2}$ in the literature. As a second application, we reproduce the correct analytical expression for a next-to-leading order (NLO) kinematic distribution from data, which is simulated with a NLO event generator. Finally, we derive analytical approximations for the NLO kinematic distributions after detector simulation, for which no known analytical formulas currently exist.

hep-ph↗

arXiv2022

Probing the Top-Higgs Yukawa CP Structure in dileptonic $t \bar t h$ with $M_2$-Assisted Reconstruction

Constraining the Higgs boson properties is a cornerstone of the LHC program. We study the potential to directly probe the Higgs-top CP-structure via the $t\bar{t}h$ channel at the LHC with the Higgs boson decaying to a bottom pair and top-quarks in the dileptonic mode. We show that a combination of laboratory and $t\bar{t}$ rest frame observables display large CP-sensitivity, exploring the spin correlations in the top decays. To efficiently reconstruct our final state, we present a method based on simple mass minimization and prove its robustness to shower, hadronization and detector effects. In addition, the mass reconstruction works as an extra relevant handle for background suppression. Based on our results, we demonstrate that the Higgs-top CP-phase $(α)$ can be probed up to $\cosα< 0.7$ at the high luminosity LHC.

hep-ph↗

arXiv2022

Portraying Double Higgs at the Large Hadron Collider II

The Higgs potential is vital to understand the electroweak symmetry breaking mechanism, and probing the Higgs self-interaction is arguably one of the most important physics targets at current and upcoming collider experiments. In particular, the triple Higgs coupling may be accessible at the HL-LHC by combining results in multiple channels, which motivates to study all possible decay modes for the double Higgs production. In this paper, we revisit the double Higgs production at the HL-LHC in the final state with two $b$-tagged jets, two leptons and missing transverse momentum. We focus on the performance of various neural network architectures with different input features: low-level (four momenta), high-level (kinematic variables) and image-based. We find it possible to bring a modest increase in the signal sensitivity over existing results via careful optimization of machine learning algorithms making a full use of novel kinematic variables.

hep-ph↗

arXiv2022

Search Prospects for Axion-like Particles at Rare Nuclear Isotope Accelerator Facilities

We propose a novel experimental scheme, called DAMSA (Dump-produced Aboriginal Matter Searches at an Accelerator), for searching for dark-sector particles, using rare nuclear isotope accelerator facilities that provide high-flux proton beams to produce a large number of rare nuclear isotopes. The high-intensity nature of their beams enables the investigation of dark-sector particles, including axion-like particles (ALPs) and dark photons. By contrast, their typical beam energies are not large enough to produce the backgrounds such as neutrinos resulting from secondary charged particles. The detector of DAMSA is then placed immediate downstream of the proton beam dump to maximize the prompt decay signals of dark-sector particles, which are often challenging to probe in other beam-dump-type experiments featuring a longer baseline, at the expense of an enormous amount of the beam-related neutron (BRN) background. We demonstrate that BRN can be significantly suppressed if the signal accompanies multiple, correlated visible particles in the final state. As an example physics case, we consider ALPs interacting with the Standard Model photon and their diphoton decay signal at DAMSA implemented at a rare nuclear isotope facility similar to the Rare isotope Accelerator complex for ON-line experiment under construction in South Korea. We show that the close proximity of the detector to the ALP production dump makes it possible to probe a high-mass region of ALP parameter space that the existing experiments have never explored.

hep-ph↗

arXiv2022

Kinematic Variables and Feature Engineering for Particle Phenomenology

Kinematic variables have been playing an important role in collider phenomenology, as they expedite discoveries of new particles by separating signal events from unwanted background events and allow for measurements of particle properties such as masses, couplings, spins, etc. For the past 10 years, an enormous number of kinematic variables have been designed and proposed, primarily for the experiments at the Large Hadron Collider, allowing for a drastic reduction of high-dimensional experimental data to lower-dimensional observables, from which one can readily extract underlying features of phase space and develop better-optimized data-analysis strategies. We review these recent developments in the area of phase space kinematics, summarizing the new kinematic variables with important phenomenological implications and physics applications. We also review recently proposed analysis methods and techniques specifically designed to leverage the new kinematic variables. As machine learning is nowadays percolating through many fields of particle physics including collider phenomenology, we discuss the interconnection and mutual complementarity of kinematic variables and machine learning techniques. We finally discuss how the utilization of kinematic variables originally developed for colliders can be extended to other high-energy physics experiments including neutrino experiments.

hep-ph↗

arXiv2022

Resolving Combinatorial Ambiguities in Dilepton $t \bar t$ Event Topologies with Neural Networks

We study the potential of deep learning to resolve the combinatorial problem in SUSY-like events with two invisible particles at the LHC. As a concrete example, we focus on dileptonic $t \bar t$ events, where the combinatorial problem becomes an issue of binary classification: pairing the correct lepton with each $b$ quark coming from the decays of the tops. We investigate the performance of a number of machine learning algorithms, including attention-based networks, which have been used for a similar problem in the fully-hadronic channel of $t\bar t$ production; and the Lorentz Boost Network, which is motivated by physics principles. We then consider the general case when the underlying mass spectrum is unknown, and hence no kinematic endpoint information is available. Compared against existing methods based on kinematic variables, we demonstrate that the efficiency for selecting the correct pairing is greatly improved by utilizing deep learning techniques.

hep-ph↗