arXiv subjects

J. Damgov

Publications and source records attributed to J. Damgov.

10 recordsLinked to original sources

arXiv2024

Vertex Imaging Hadron Calorimetry Using AI/ML Tools

The fluctuations in energy loss to processes that do not generate measurable signals, such as binding energy losses, set the limit on achievable hadronic energy resolution in traditional energy reconstruction techniques. The correlation between the number of hadronic interaction vertices in a shower and invisible energy is found to be strong and is used to estimate invisible energy fraction in highly granular calorimeters in short time intervals (<10 ns). We simulated images of hadronic showers using GEANT4 and deployed a neural network to analyze the images for energy regression. The neural network-based approach results in significant improvement in energy resolution, from 13 % to 4 % in the case of a Cherenkov calorimeter for 100 GeV pion showers. We discuss the significance of the phenomena responsible for this improvement and the plans for experimental verification of these results and further development.

physics.ins-det↗

arXiv2024

High-granularity Dual-readout Calorimeter: Evolution of a Classic Prototype

The original dual-readout calorimeter prototype (DREAM), constructed two decades ago, has proven instrumental in advancing our understanding of calorimetry. It has facilitated a multitude of breakthroughs by leveraging signals from complementary media (Cherenkov and scintillation) to capture fluctuations in electromagnetic energy fraction within hadronic showers. Over the years, extensive studies have shed light on the performance characteristics of this module, rendering it exceptionally well-understood. Drawing on this wealth of experience, we have embarked on enhancing the detectors' capabilities further by integrating fast silicon photomultipliers (SiPMs) with finer transverse segmentation, $\sim$1 cm$^2$, as well as longitudinal segmentation by timing measuring better than 10 cm. This configuration will allow us to image hadronic showers with high granularity (HG-DREAM). We argue that the spatial information provided by such a granular detector in a short time window ($\approx$5 ns) leads to substantial enhancement in energy resolution when advanced neural networks are employed in energy reconstruction. We briefly present the current status of work, new concepts that have been introduced to the detector, and expectations from simulations.

physics.ins-det↗

arXiv2022

Perspectives on the Calibration of CNN Energy Reconstruction in Highly Granular Calorimeters

We present a study which shows encouraging stability of the response linearity for a simulated high granularity calorimeter module reconstructed by a CNN model to miscalibration, bias, and noise effects. Our results also show an intuitive, quantifiable relationship between these factors and the calibration parameters. We trained a CNN model to reconstruct energy in the calorimeter module using simulated single-pion events; we then observed the response of the model under various miscalibration, bias, and noise conditions that affected the model input. From these data, we estimated linear response models to calibrate the CNN. We also quantified the relationship between these factors and the calibration parameters by regression analysis.

physics.ins-det↗

arXiv2022

Deep learning applications for quality control in particle detector construction

The growing complexity of particle detectors makes their construction and quality control a new challenge. We present studies that explore the use of deep learning-based computer vision techniques to perform quality checks of detector components and assembly steps, which will automate procedures and minimize the need for human interventions. This study focuses on the construction steps of a silicon detector, which involve forming a mechanical structure with the sensor and wire bonding individual cells to electronics for reading out signals. Silicon detectors in high energy physics experiments today have millions of channels. Manual quality control of these and other high channel-density detectors requires enormous amounts of labor and can be prone to errors. Here, we explore computer vision applications to either augment or fully replace visual inspections done by humans. We investigated convolutional neural networks for image classification and autoencoders for anomalies detection. Two proof-of-concept studies will be presented.

hep-ex↗

arXiv2022

On the Use of Neural Networks for Energy Reconstruction in High-granularity Calorimeters

We contrasted the performance of deep neural networks - Convolutional Neural Network (CNN) and Graph Neural Network (GNN) - to current state of the art energy regression methods in a finely 3D-segmented calorimeter simulated by GEANT4. This comparative benchmark gives us some insight to assess the particular latent signals neural network methods exploit to achieve superior resolution. A CNN trained solely on a pure sample of pions achieved substantial improvement in the energy resolution for both single pions and jets over the conventional approaches. It maintained good performance for electron and photon reconstruction. We also used the Graph Neural Network (GNN) with edge convolution to assess the importance of timing information in the shower development for improved energy reconstruction. We implement a simple simulation based correction to the energy sum derived from the fraction of energy deposited in the electromagnetic shower component. This serves as an approximate dual-readout analogue for our benchmark comparison. Although this study does not include the simulation of detector effects, such as electronic noise, the margin of improvement seems robust enough to suggest these benefits will endure in real-world application. We also find reason to infer that the CNN/GNN methods leverage latent features that concur with our current understanding of the physics of calorimeter measurement.

physics.ins-det↗

arXiv2019

Cerium-Doped Fused-Silica Fibers as Wavelength Shifters

We have evaluated the performance of a Ce-doped fused-silica fiber as wavelength shifter coupled to a CeF$_{3}$ crystal using electron beams at CERN. The pulse shape and collection efficiency were measured using irradiated (100 kGy) and un-irradiated fibers. In addition, we evaluated the light yield of various Ce-doped fibers and explored the possibility of using them in the future, including for precision timing applications in a high-luminosity collider environment.

physics.ins-det↗

arXiv2007

Search for a Standard Model Higgs Boson in CMS via Vector Boson Fusion in the H->WW->lνlνChannel

We present the potential for discovering the Standard Model Higgs boson produced by the vector-boson fusion mechanism. We considered the decay of Higgs bosons into the W+W- final state, with both W-bosons subsequently decaying leptonically. The main background is ttbar with one or more jets produced. This study is based on a full simulation of the CMS detector, and up-to-date reconstruction codes. The result is that a signal of 5 sigma significance can be obtained with an integrated luminosity of 12-72 1/fb for Higgs boson masses between 130-200 GeV. In addition, the major background can be measured directly to 7% from the data with an integrated luminosity of 30 1/fb. In this study, we also suggested a method to obtain information in Higgs mass using the transverse mass distributions.

hep-ex↗

arXiv2001

Simultaneously non-linear energy calibration of CMS calorimeters for single pions and electrons

CMS calorimeter energy calibration was done in the full CMS simulated geometry for the pseudorapidity region eta = 0. The samples of single pion events were generated with a set of incident energies from 5 GeV to 3 TeV and for single electrons from 5 to 500 GeV. The analysis of the simulated data shows that standard calibration using just sampling coefficients for calorimeter parts with different sampling ratio gives nonlinear calorimeter response. Non-linear calibration technique was applied simultaneously for pion and electron beams which is preparation for jets energy reconstruction. It improve calorimeter energy resolution for pions and restore the calorimeter linearity.

hep-ex↗

arXiv2001

Non-linear energy calibration of CMS calorimeters for single pions

CMS calorimeter energy calibration was done in the full CMS simulated geometry for the pseudorapidity region eta = 0. The samples of single pion events were generated with a set of incident energies from 10 GeV to 3 TeV. The analysis of the simulated data shows that standard calibration using just sampling coefficients for calorimeter parts with different sampling ratio gives nonlinear calorimeter response. Non-linear calibration technique was applied for improving calorimeter energy resolution and restoring the calorimeter linearity.

hep-ex↗

arXiv2000

Application of Neural Networks for Energy Reconstruction

The possibility to use Neural Networks for reconstruction of the energy deposited in the calorimetry system of the CMS detector is investigated. It is shown that using feed - forward neural network, good linearity, Gaussian energy distribution and good energy resolution can be achieved. Significant improvement of the energy resolution and linearity is reached in comparison with other weighting methods for energy reconstruction.

hep-ex↗