A5020033662- Particle physics theoretical and experimental studies
- High-Energy Particle Collisions Research
- Quantum Chromodynamics and Particle Interactions
- Particle Detector Development and Performance
- Computational Physics and Python Applications
- Dark Matter and Cosmic Phenomena
- Cosmology and Gravitation Theories
- Neutrino Physics Research
- Distributed and Parallel Computing Systems
- Advanced Data Storage Technologies
- Medical Imaging Techniques and Applications
- Parallel Computing and Optimization Techniques
- Black Holes and Theoretical Physics
- Scientific Computing and Data Management
- Astrophysics and Cosmic Phenomena
- Radiation Detection and Scintillator Technologies
- Gamma-ray bursts and supernovae
- Atomic and Subatomic Physics Research
- Graph Theory and Algorithms
- Stochastic processes and financial applications
- Advanced Graph Neural Networks
- 3D Shape Modeling and Analysis
- Particle Accelerators and Free-Electron Lasers
- Optical properties and cooling technologies in crystalline materials
- Nuclear physics research studies
Fermi National Accelerator Laboratory
2019-2025
Institute of High Energy Physics
2022-2024
A. Alikhanyan National Laboratory
2022-2024
University of Antwerp
2024
University of Nebraska–Lincoln
2022-2023
University of Florida
2023
ETH Zurich
2017-2019
Pattern recognition problems in high energy physics are notably different from traditional machine learning applications computer vision. Reconstruction algorithms identify and measure the kinematic properties of particles produced collisions recorded with complex detector systems. Two critical reconstruction charged particle trajectories tracking detectors showers calorimeters. These two have unique challenges characteristics, but both dimensionality, degree sparsity, geometric layouts....
The observation of Higgs boson production in association with a top quark-antiquark pair is reported, based on combined analysis proton-proton collision data at center-of-mass energies √s = 7, 8, and 13 TeV, corresponding to integrated luminosities up 5.1, 19.7, 35.9 fb^(-1), respectively. were collected the CMS detector CERN LHC. results statistically independent searches for bosons produced conjunction decaying pairs W bosons, Z photons, τ leptons, or bottom quark jets are maximize...
The Exa.TrkX project has applied geometric learning concepts such as metric and graph neural networks to HEP particle tracking. Exa.TrkX's tracking pipeline groups detector measurements form track candidates filters them. pipeline, originally developed using the TrackML dataset (a simulation of an LHC-inspired detector), been demonstrated on other detectors, including DUNE Liquid Argon TPC CMS High-Granularity Calorimeter. This paper documents new developments needed study physics computing...
This document provides a writeup of all contributions to the workshop on "High precision measurements $α_s$: From LHC FCC-ee" held at CERN, Oct. 12--13, 2015. The explored in depth latest developments determination QCD coupling $α_s$ from 15 methods where high are (or will be) available. Those include low-energy observables: (i) lattice QCD, (ii) pion decay factor, (iii) quarkonia and (iv) $τ$ decays, (v) soft parton-to-hadron fragmentation functions, as well high-energy (vi) global fits...
Abstract In the next decade, demands for computing in large scientific experiments are expected to grow tremendously. During same time period, CPU performance increases will be limited. At CERN Large Hadron Collider (LHC), these two issues confront one another as collider is upgraded high luminosity running. Alternative processors such graphics processing units (GPUs) can resolve this confrontation provided that algorithms sufficiently accelerated. many cases, algorithmic speedups found...
Abstract Recently, graph neural networks (GNNs) have been successfully used for a variety of particle reconstruction problems in high energy physics, including tracking. The Exa.TrkX pipeline based on GNNs demonstrated promising performance reconstructing tracks dense environments. It includes five discrete steps: data encoding, building, edge filtering, GNN, and track labeling. All steps were written Python run both GPUs CPUs. In this work, we accelerate the implementation through...
Abstract We present the current stage of research progress towards a one-pass, completely Machine Learning (ML) based imaging calorimeter reconstruction. The model used is on Graph Neural Networks (GNNs) and directly analyzes hits in each HGCAL endcap. ML algorithm trained to predict clusters originating from same incident particle by labeling with cluster index. impose simple criteria assess whether associated as prediction are matched those resulting any particular individual particles....
Computing needs for high energy physics are already intensive and expected to increase drastically in the coming years. In this context, heterogeneous computing, specifically as-a-service has potential significant gains over traditional computing models. Although previous studies packages field of have focused on GPUs as accelerators, FPGAs an extremely promising option well. A series workflows developed establish performance capabilities a service. Multiple different devices range...
To address the unprecedented scale of HL-LHC data, Exa.TrkX project is investigating a variety machine learning approaches to particle track reconstruction. The most promising these solutions, graph neural networks (GNN), process event as that connects measurements (detector hits corresponding nodes) with candidate line segments between (corresponding edges). Detector information can be associated nodes and edges, enabling GNN propagate embedded parameters around predict node-, edge-...
We present a determination of the strong coupling constant αsmZ using inclusive top-quark pair production cross section measurements performed at LHC and Tevatron. Following procedure first applied by CMS Collaboration, we extract individual values from different experiments several centre-of-mass energies, QCD predictions complete in NNLO perturbation theory, supplemented with NNLL approximations to all orders, suitable sets parton distribution functions. The determinations are then...
This paper presents a graph neural network (GNN) technique for low-level reconstruction of neutrino interactions in Liquid Argon Time Projection Chamber (LArTPC). GNNs are still relatively novel technique, and have shown great promise similar tasks the LHC. In this paper, multihead attention message passing is used to classify relationship between detector hits by labelling edges, determining whether were produced same underlying particle, if so, particle type. The trained model 84% accurate...
Abstract The Exa.TrkX project presents a graph neural network (GNN) technique for low-level reconstruction of neutrino interactions in Liquid Argon Time Projection Chamber (LArTPC). GNNs are still relatively novel technique, and have shown great promise similar tasks the Large Hadron Collider (LHC). Graphs describing particle formed by treating each detector hit as node, with edges relationships between hits. We utilise multi-head attention message passing which performs convolutions order...
Classifying whole images is a classic problem in machine learning, and graph neural networks are powerful methodology to learn highly irregular geometries. It often the case that certain parts of point cloud more important than others when determining overall classification. On structures this started by pooling information at end convolutional filters, has evolved variety staged techniques on static graphs. In paper, dynamic formulation introduced removes need for predetermined structure....
Particle tracking is a challenging pattern recognition task at the Large Hadron Collider (LHC) and High Luminosity-LHC. Conventional algorithms, such as those based on Kalman Filter, achieve excellent performance in reconstructing prompt tracks from collision points. However, they require dedicated configuration additional computing time to efficiently reconstruct large radius created away We developed an end-to-end machine learning-based track finding algorithm for HL-LHC, Exa.TrkX...
In this community review report, we discuss applications and techniques for fast machine learning (ML) in science -- the concept of integrating power ML methods into real-time experimental data processing loop to accelerate scientific discovery. The material report builds on two workshops held by Fast Science covers three main areas: across a number domains; training implementing performant resource-efficient algorithms; computing architectures, platforms, technologies deploying these...