Search Results (21)

The proliferation of sensors brings an immense volume of spatio-temporal (ST) data in many domains, including monitoring, diagnostics, and prognostics applications. Data curation is a time-consuming process for a large volume of data, making it challenging and expensive to deploy data analytics platforms in new environments. Transfer learning (TL) mechanisms promise to mitigate data sparsity and model complexity by utilizing pre-trained models for a new task. Despite the triumph of TL in fields like computer vision and natural language processing, efforts on complex ST models for anomaly detection (AD) applications are limited. In this study, we present the potential of TL within the context of high-dimensional ST AD with a hybrid autoencoder architecture, incorporating convolutional, graph, and recurrent neural networks. Motivated by the need for improved model accuracy and robustness, particularly in scenarios with limited training data on systems with thousands of sensors, this research investigates the transferability of models trained on different sections of the Hadron Calorimeter of the Compact Muon Solenoid experiment at CERN. The key contributions of the study include exploring TL’s potential and limitations within the context of encoder and decoder networks, revealing insights into model initialization and training configurations that enhance performance while substantially reducing trainable parameters and mitigating data contamination effects. Full article

The civil engineering industry faces a critical need for innovative non-destructive evaluation methods, particularly for ageing critical infrastructure, such as bridges, where current techniques fall short. Muography, a non-invasive imaging technique, constructs three-dimensional density maps by detecting the interactions of naturally occurring cosmic-ray muons within the scanned volume. Cosmic-ray muons offer both deep penetration capabilities due to their high momenta and inherent safety due to their natural source. However, the technology’s reliance on this natural source results in a constrained muon flux, leading to prolonged acquisition times, noisy reconstructions, and challenges in image interpretation. To address these limitations, we developed a two-model deep learning approach. First, we employed a conditional Wasserstein Generative Adversarial Network with Gradient Penalty (cWGAN-GP) to perform predictive upsampling of undersampled muography images. Using the Structural Similarity Index Measure (SSIM), 1-day sampled images were able to match the perceptual qualities of a 21-day image, while the Peak Signal-to-Noise Ratio (PSNR) indicated a noise improvement to that of 31 days worth of sampling. A second cWGAN-GP model, trained for semantic segmentation, was developed to quantitatively assess the upsampling model’s impact on each of the features within the concrete samples. This model was able to achieve segmentation of rebar grids and tendon ducts embedded in the concrete, with respective Dice–Sørensen accuracy coefficients of 0.8174 and 0.8663. This model also revealed an unexpected capability to mitigate—and in some cases entirely remove—z-plane smearing artifacts caused by the muography’s inherent inverse imaging problem. Both models were trained on a comprehensive dataset generated through Geant4 Monte Carlo simulations designed to reflect realistic civil infrastructure scenarios. Our results demonstrate significant improvements in both acquisition speed and image quality, marking a substantial step toward making muography more practical for reinforced concrete infrastructure monitoring applications. Full article

The CREDO-Maze project is the concept for a network of stations recording local, extensive cosmic ray air showers. Each station consists of four small scintillation detectors and a control unit that monitors the cosmic ray flux 24 h a day and transmits the results to the central server. The modular design of each array allows the results to be used in educational classes on nuclear radiation, relativistic physics, and particle physics and as a teaching aid in regular school classrooms and more. As an example, we present here some preliminary results from the CREDO-Maze muon telescope missions to the Arctic and down into a deep salt mine, as well as the first shower-particle correlation measurements from a table-top experiment at Walailak University. These experiments show that the different geometric configurations of the CREDO-Maze detector set can be used for projects beyond the scope of the secondary school curriculum, and they can form the basis of student theses and dissertations at universities. Full article

The High-Luminosity Large Hadron Collider (HL-LHC) project aims to improve the performance of the LHC by increasing the proton–proton collision luminosity. New physics discoveries will be possible starting in 2027. The HL-LHC aims to improve the integrated luminosity by a factor of 10 concerning the current running LHC’s design value. The HL-LHC project foresees delivering proton–proton collisions at 14 TeV CM (Center of Mass) energy providing the integrated luminosity to a value of 3 ab⁻¹ for the ATLAS and CMS experiments, 50 fb⁻¹ for LHCb, and 5 fb⁻¹ for ALICE. The increased integrated luminosity for the above LHC experiments will provide the potential to discover rare processes while improving these measurements’ signal-to-noise (S/N) ratio statistics. The ATLAS muon spectrometer has been upgraded to face the challenges of the luminosity at the HL-LHC run. The new sub-detectors are as follows: The New Small Wheel (NSW) has replaced the Cathode Strip Chambers (CSC) discs at the internal part of the ATLAS end cups. The new integrated small Monitored Drift Chambers (sMDT) with the Resistive Plate Chambers (RPC) are installed at the outer end of the ATLAS BI (Barrel Inner) layer, in the barrel–endcap transition region, at 1.0 < |η| < 1.3, where η is the pseudo-rapidity (pseudo-rapidity η is a commonly used spatial coordinate describing the angle of a particle relative to the beam axis, defined as $\eta =-ln\left[tan\left({\displaystyle \frac{\theta }{2}}\right)\right]$, where θ is the angle between the vector momentum $\overrightarrow{p}$ and the positive direction of the beam axis). The NSW is an innovative technological achievement, including the MicroMegas (MM) gas detectors in large areas and small-strip Thin Gap Chambers (sTGC), enabling high p_T (high p_T is the high value of the particles’ transverse momentum versus the beam collision axis) trigger and muon detection. The muon reconstruction, the background rate, other spectrometer parameters, and the NSW performance are also presented. Full article

Following the decommissioning of the Main Injector Neutrino Oscillation Search (MINOS) experiment, muon and hadron monitors have emerged as vital diagnostic tools for the NuMI Off-axis ${\nu }_{\mu }$ Appearance (NOvA) experiment at Fermilab. These tools are crucial for overseeing the Neutrinos at the Main Injector (NuMI) beam. This study endeavors to ensure the monitor signal quality and to correlate them with the Neutrino beam profile. Leveraging muon monitor simulations, we systematically explore the monitor responses to variations in proton-beam and lattice parameters. Through the amalgamation of individual pixel data from muon monitors, pattern-recognition algorithms, simulations, and measured data, we devise machine-learning-based models to predict muon monitor responses and Neutrino flux. Full article

The Compact Muon Solenoid (CMS) experiment is a general-purpose detector for high-energy collision at the Large Hadron Collider (LHC) at CERN. It employs an online data quality monitoring (DQM) system to promptly spot and diagnose particle data acquisition problems to avoid data quality loss. In this study, we present a semi-supervised spatio-temporal anomaly detection (AD) monitoring system for the physics particle reading channels of the Hadron Calorimeter (HCAL) of the CMS using three-dimensional digi-occupancy map data of the DQM. We propose the GraphSTAD system, which employs convolutional and graph neural networks to learn local spatial characteristics induced by particles traversing the detector and the global behavior owing to shared backend circuit connections and housing boxes of the channels, respectively. Recurrent neural networks capture the temporal evolution of the extracted spatial features. We validate the accuracy of the proposed AD system in capturing diverse channel fault types using the LHC collision data sets. The GraphSTAD system achieves production-level accuracy and is being integrated into the CMS core production system for real-time monitoring of the HCAL. We provide a quantitative performance comparison with alternative benchmark models to demonstrate the promising leverage of the presented system. Full article

In the search for the CP violation (CPV) in the leptonic sector, crucial information was obtained a decade ago from reactor and accelerator experiments. The discovery and measurement of the third neutrino mixing angle, ${\theta }_{13}$, with a value ∼9${}^{\circ }$, allow for the possibility to discover the leptonic Dirac CP-violating angle, ${\delta }_{CP}$, with long baseline neutrino Super Beams. ESS$\nu$SB is a long-baseline neutrino project that will be able to measure the CPV in the leptonic sector at the second oscillation maximum, where the sensitivity of the experiment is higher compared to that at the first oscillation maximum. The extension project, ESS$\nu$SB+, aims to address a very challenging task on measuring the neutrino–nucleon cross-section, which is the dominant term of the systematic uncertainty, in the energy range 0.2–0.6 GeV, using a Low-Energy nuSTORM (LEnuSTORM) and an ENUBET-like Low-Energy Monitored Neutrino Beam (LEMNB) facilities. The target station plays the main role in generating a well defined and focused pion, and hence muon, beam. Full article

The NuMI target facility at Fermilab produces an intense muon neutrino beam for the NOvA (NuMI Off-axis ${\nu }_e$ Appearance) long baseline neutrino experiment. Three arrays of muon monitors located downstream of the hadron absorber in the NuMI beamline provide the measurements of the primary beam and horn current quality. We have studied the response of muon monitors with the proton beam profile changes and focusing horn current variations. The responses of muon monitors are used to develop machine learning (ML) algorithms to monitor the beam quality. We present the development of the machine learning applications and future plans. This effort is important for future applications such as beam quality assurance, anomaly detection, and neutrino beam systematics studies. Our results demonstrate the advantages of developing useful ML applications that can be leveraged for future beamlines such as LBNF. Full article

Monitored neutrino beams are facilities where beam diagnostics enable the counting and identification of charged leptons in the decay tunnel of a narrow band beam. These facilities can monitor neutrino production at the single particle level (flux precision $<1$%) and provide information about the neutrino energy at the 10% level. The ENUBET Collaboration has demonstrated that lepton monitoring might be achieved not only by employing kaon decays but also by identifying muons from the ${\pi }^{+}\to {\mu }^{+}{\nu }_{\mu }$ decays and positrons from the decay-in-flight of muons before the hadron dump. As a consequence, beam monitoring can be performed using the ENUBET technique even when the kaon production yield is kinematically suppressed. This finding opens up a wealth of opportunities for measuring neutrino cross-sections below 1 GeV. In this paper, we investigate this opportunity at the European Spallation Source (ESS), which is an ideal facility to measure ${\nu }_{\mu }$ and ${\nu }_e$ cross-sections in the 0.2–1 GeV range. We also describe the planned activities for the design of this beam at the ESS within the framework of the ESS$\nu$SB+ design study, which was approved by the EU in July 2022. Full article

The main source of systematic uncertainty on neutrino cross-section measurements at the GeV scale originates from the poor knowledge of the initial flux. The goal of reducing this uncertainty to 1% can be achieved through the monitoring of charged leptons produced in association with neutrinos, by properly instrumenting the decay region of a conventional narrow-band neutrino beam. Large-angle muons and positrons from kaons are measured by a sampling calorimeter on the decay tunnel walls, while muon stations after the hadron dump can be used to monitor the neutrino component from pion decays. Furthermore, the narrow momentum width (<10%) of the beam provides a O (10%) measurement of the neutrino energy on an event-by-event basis, thanks to its correlation with the radial position of the interaction at the neutrino detector. The ENUBET project has been funded by the ERC in 2016 to prove the feasibility of such a monitored neutrino beam and, since 2019, ENUBET is also a CERN neutrino platform experiment (NP06/ENUBET). The breakthrough the project achieved is the design of a horn-less neutrino beamline that would allow for a 1% measurement of ${\nu }_e$ and ${\nu }_{\mu }$ cross-sections in about 3 years of data taking at CERN-SPS, using ProtoDUNE as far detector. Full article

High-energy physics in the atmosphere (HEPA) has undergone an intense reformation in the last decade. Correlated measurements of particle fluxes modulated by strong atmospheric electric fields, simultaneous measurements of the disturbances of the near-surface electric fields and lightning location, and registration of various meteorological parameters on the Earth have led to a better understanding of the complex processes in the terrestrial atmosphere. The cooperation of cosmic rays and atmospheric physics has led to the development of models for the origin of particle bursts recorded on the Earth’s surface, estimation of vertical and horizontal profiles of electric fields in the lower atmosphere, recovery of electron and gamma ray energy spectra, the muon deceleration effect, etc. The main goal of this review is to demonstrate how the measurements performed at the Aragats cosmic ray observatory led to new results in atmospheric physics. We monitored particle fluxes around the clock using synchronized networks of advanced sensors that recorded and stored multidimensional data in databases with open, fast, and reliable access. Visualization and statistical analysis of particle data from hundreds of measurement channels disclosed the structure and strength of the atmospheric electric fields and explained observed particle bursts. Consequent solving of direct and inverse problems of cosmic rays revealed the modulation effects that the atmospheric electric field has on cosmic ray fluxes. Full article

Aging infrastructure is a threatening issue throughout the world. Long exposure to oxygen and moisture causes premature corrosion of reinforced concrete structures leading to the collapse of the structures. As a consequence, real-time monitoring of civil structures for rust becomes critical in avoiding mishaps. Muon scattering tomography is a non-destructive, non-invasive technique which has shown impressive results in 3D imaging of civil structures. This paper explores the application of advanced machine learning techniques in identifying a rusted reinforced concrete rebar using muon scattering tomography. To achieve this, we have simulated the performance of an imaging prototype setup, designed to carry out muon scattering tomography, to precisely measure the rust percentage in a rusted rebar. We have produced a 2D image based on the projected 3D scattering vertices of the muons and used the scattering vertex density and average deviation angle per pixel as the distinguishing parameter for the analysis. A filtering algorithm, namely the Pattern Recognition Method, has been employed to eliminate background noise. Since this problem boils down to whether or not the material being analyzed is rust, i.e., a classification problem, we have adopted the well-known machine learning algorithm Support Vector Machine to identify rust in the rusted reinforced cement concrete structure. It was observed that the trained model could easily identify 30% of rust in the structure with a nominal exposure of 30 days within a small error range of 7.3%. Full article

The Tile Calorimeter (TileCal) is the central hadronic calorimeter of the ATLAS experiment at the LHC. This sampling device is made of steel plates acting as absorber and scintillating tiles as active medium. The wavelength-shifting fibers collect the light from scintillators and carry it to the photomultiplier tubes (PMTs). The analogue signals from the PMTs are amplified, shaped and digitized by sampling the signal every 25 ns and stored on detector until a trigger decision is received. The TileCal front-end electronics read out the signals produced by 9852 channels, whose dynamic range covers the interval from 30 MeV to 2 TeV. Each stage of the signal propagation from scintillation light to the signal reconstruction is monitored and calibrated. During LHC Run-2, high-momentum isolated muons and isolated hadrons have been used to study and validate the electromagnetic scale and the hadronic response, respectively. The time resolution was studied with multi-jet events. Results of performance studies that address calibration, stability, energy scale, uniformity and time resolution are presented. Full article

Muon tomography or muography is an innovative imaging technique using atmospheric muons. The technique is based on the detection of muons that have crossed a target and the measurement of their attenuation or deviation induced by the medium. Muon flux models are key ingredients to convert tomographic and calibration data into the 2D or 3D density maps of the target. Ideally, they should take into account all possible types of local effects, from geomagnetism to atmospheric conditions. Two approaches are commonly used: semi-empirical models or Monte Carlo simulations. The latter offers the advantage to tackle down many environmental and experimental parameters and also allows the optimization of the nearly horizontal muons flux, which remains a long-standing problem for many muography applications. The goal of this paper is to identify through a detailed simulation what kind of environmental and experimental effects may affect the muography imaging sensitivity and its monitoring performance. The results have been obtained within the CORSIKA simulation framework, which offers the possibility to tune various parameters. The paper presents the simulation’s configuration and the results obtained for the muon fluxes computed in various conditions. Full article

In this study, we analyze the L3 precision muon spectrometer data from November 2000. The results showed that a 4.7σ muon excess appeared at a time coincident with the solar flare of 8 November 2000. This muon excess corresponded to primary protons above 40 GeV, coming from a sky cell of solid angle 0.048 sr. The probability of being a background fluctuation was estimated to be about 0.1%. It is interesting and noteworthy that an M-class solar flare may also accelerate solar protons to such high energies. Full article