arXiv subjects

Satyaki Bhattacharya

Publications and source records attributed to Satyaki Bhattacharya.

16 recordsLinked to original sources

arXiv2026

Enhancing Electromagnetic Calorimeter Signal Reconstruction with Machine Learning-Based Noise Discrimination

Calorimeters operating in high-radiation environments are susceptible to damage, leading to increased noise that can significantly degrade energy resolution. A common way to mitigate noise is to apply a higher energy threshold on the calorimeter cells, typically set a few standard deviations above the noise level. However, this method risks discarding cells with genuine energy deposits, worsening the energy resolution and the energy deposit pattern. In this paper, we investigate graph neural network (GNN) based algorithms as an alternative to rigid energy thresholds. The proposed approach exploits the full pulse-shape information together with the correlations among energy deposits in individual cells within an electromagnetic cluster. The study is performed using a standalone Geant4 simulation of an 11x11 matrix of lead tungstate crystals. To simulate it in more relatic way We demonstrate that the machine learning based method significantly outperforms a simple threshold based strategy, achieving an improvement in the calorimeter energy resolution of up to 82% and recovering up to 83% of true signal cells within the calorimeter grid.

physics.ins-det↗

arXiv2026

PySiPMGUI: A Universal Python-Based Software for Photodetector I-V Quality Assurance: From Underground Dark Matter Searches to Astroparticle Cherenkov Cameras

Silicon photomultipliers (SiPMs) are currently the most prevalent photon detection technology in modern experiments in high-energy physics, astroparticle physics, neutrino physics, and dark matter searches. The high detection efficiency for photons, excellent timing resolution, small size, and magnetic field independence make them ideal for precision measurements in low-light conditions. However, key parameters like breakdown voltage, gain, and dark count rate show a strong dependence on the bias voltage and temperature, requiring a systematic characterization. In this work, we present an open-source graphical user interface (GUI) for automated SiPM characterization, leveraging standard laboratory instrument communication protocols (PyVISA). The tool provides a free, open, and platform-independent solution for detector R&D and large-scale SiPM characterization, and is available at this GitHuB repository [1]. The analysis procedure is validated against the manufacturer's datasheet for the SensL MicroFC-60035-SMT, yielding results in good agreement for both breakdown voltage and dark count rate.

physics.ins-det↗

arXiv2026

Two-dimensional Rademacher walk

We study a generalisation of the one-dimensional Rademacher random walk introduced in Bhattacharya and Volkov (2023) to $\mathbb{Z}^2$ (for $d\ge 3$, the Rademacher random walk is always transient, as follows from Theorem 8.8 in Englander and Volkov (2025)). This walk is defined as the sum of a sequence of independent steps, where each step goes in one of the four possible directions with equal probability, and the size of the $n$th step is $a_n$ where $\{a_n\}$ is a given sequence of positive integers. We establish some general conditions under which the walk is recurrent or transient.

math.PR↗

arXiv2025

Recurrence, transience and anti-concentration of Rademacher random walks

The Rademacher random walk associated with a deterministic sequence $(a_n)_{n \geq 1}$ is the walk which starts at zero and, at step $i$, independently steps either up or down by $a_i$ with equal probability. We continue the study begun by Bhattacharya and Volkov in 2023 of the transience or recurrence of one-dimensional Rademacher random walks. In particular, we show that if the sequence of step sizes is bounded, the walk is weakly recurrent, meaning that it returns infinitely often to a random finite interval, while if the step sizes tend to infinity arbitrarily slowly, the walk may be transient. On the other hand, using a construction with integer step sizes, we show that the step sizes may grow arbitrarily fast and still give a weakly recurrent random walk. We also show, using a construction with non-integer step sizes, that the same conclusion holds even if we restrict to strictly increasing step sizes. However, we prove that if $a_n = n^{α+ o(1)}$ for some $α> 1/2$, then the walk is transient. We show that the bound on $α$ is tight by giving an example where $a_n = Θ(n^{1/2})$ and the walk is weakly recurrent.

math.PR↗

arXiv2025

Diphoton signals for the Georgi-Machacek scenario at the Large Hadron Collider

The diphoton channel for exploring the Georgi-Machacek (GM) scenario containing scalar triplets at the Large Hadron Collider (LHC) has been identified as germane, and subjected to a detailed study. The scalar spectrum of the model, which imposes a custodial SU(2) on the potential, gets classified into a 5-plet, a 3-plet and two singlets under the custodial symmetry. While most attempts to probe or constrain the scenario at the LHC depend largely on signals of charged scalars, we point out that the custodial SU(2) singlet state H can have a substantial branching ratio (amounting to a few percent) into two photons. We carry out a detailed simulation of the resulting signal and the standard model backgrounds, obtaining the signal significance in different regions of the parameter space using the profile likelihood ratio method. Substantial regions of the GM parameter space is thus shown to be accessible to LHC studies, both at the high-luminosity run with $\int {\cal L} dt = 3000 fb^{-1}$, and also in Run-3 with $\int {\cal L} dt = 300 fb^{-1}$, even after folding in systematic errors. We have also demonstrated that a rather subtantial improvement in the signal significance is achieved by switching over from a cut-based analysis to one based on neural network.

hep-ph↗

arXiv2025

Forest Fire Model on $\mathbb{Z}_{+}$ with Delays

We consider a generalization of the forest fire model on $\mathbb{Z}_+$ with ignition at zero only, studied in [arXiv:0907.1821]. Unlike that model, we allow delays in the spread of the fires as well as the non-zero burning time of individual ``trees''. We obtain some general properties for this model, which cover, among others, the phenomena of an ``infinite fire'', not present in the original model.

math.PR↗

arXiv2024

Transience of continuous-time conservative random walks

We consider two continuous-time generalizations of conservative random walks introduced in [J.Englander and S.Volkov (2022)], an orthogonal and a spherically-symmetrical one; the latter model is known as {\em random flights}. For both models, we show the transience of the walks when $d\ge 2$ and the rate of changing of direction follows power law $t^{-α}$, $0<α\le 1$, or the law $(\ln t)^{-β}$ where $β>2$.

math.PR↗

arXiv2024

Pulse Shape Simulation and Discrimination using Machine-Learning Techniques

An essential metric for the quality of a particle-identification experiment is its statistical power to discriminate between signal and background. Pulse shape discrimination (PSD) is a basic method for this purpose in many nuclear, high-energy and rare-event search experiments where scintillation detectors are used. Conventional techniques exploit the difference between decay-times of the pulses from signal and background events or pulse signals caused by different types of radiation quanta to achieve good discrimination. However, such techniques are efficient only when the total light-emission is sufficient to get a proper pulse profile. This is only possible when adequate amount of energy is deposited from recoil of the electrons or the nuclei of the scintillator materials caused by the incident particle on the detector. But, rare-event search experiments like direct search for dark matter do not always satisfy these conditions. Hence, it becomes imperative to have a method that can deliver a very efficient discrimination in these scenarios. Neural network based machine-learning algorithms have been used for classification problems in many areas of physics especially in high-energy experiments and have given better results compared to conventional techniques. We present the results of our investigations of two network based methods \viz Dense Neural Network and Recurrent Neural Network, for pulse shape discrimination and compare the same with conventional methods.

physics.ins-det↗

arXiv2022

Recurrence and transience of Rademacher series

We introduce the notion of {\bf a}-walk $S(n)=a_1 X_1+\dots+a_n X_n$, based on a sequence of positive numbers ${\bf a}=(a_1,a_2,\dots)$ and a Rademacher sequence $X_1,X_2,\dots$. We study recurrence/transience (properly defined) of such walks for various sequences of ${\bf a}$. In particular, we establish the classification in the cases where $a_k=\lfloor k^β\rfloor$, $β>0$, as well as in the case $a_k=\lceil \log_γk \rceil$ or $a_k=\log_γk$ for $γ>1$.

math.PR↗

arXiv2020

The mono-Higgs + MET signal at the Large Hadron Collider: a study on the $γγ$ and $b\bar{b}$ final states

We investigate the potential of the channel {\em mono-Higgs + MET} in yielding signals of dark mater at the high-luminosity Large Hadron Collider (LHC). As illustration, a scalar dark matter in a Higgs portal scenario has been chosen, whose phenomenological viability has been ensured by postulating the existence of dimension-6 operators that enable cancellation in certain amplitudes for elastic scattering of dark matter in direct search experiments. These operators are found to have non-negligible contribution to the mono-Higgs signal. Thereafter, we carry out a detailed analysis of this signal, with the accompanying MET providing a useful handle in suppressing backgrounds. Signals for the Higgs decaying into both the diphoton and $b{\bar b}$ channels have been studied. A cut-based simulation is presented first, followed by a demonstration of how the statistical significance can be improved through analyses based on Boosted Decision Trees and Artificial Neural Network. The improvement is found to be especially noticeable for the $b{\bar b}$ channel.

hep-ph↗

arXiv2020

Performance Studies of the p-spray/p-stop implanted Si Sensors for the SiD Detector

Silicon Detector (SiD) is one of the proposed detector for future $e^+e^-$ Linear colliders, like International Linear Collider (ILC). The estimated neutron background for ILC is around 1 - 1.6 x 1010 1-MeV equivalent neutrons cm-2 year-1 for the Si micro strip sensors to be used in the innermost vertex detector. The p+n-n+ double-sided Si strip sensors are supposed to be used as position sensitive sensors for SiD. On the $n^+n^-$ side of these sensors, shorting due to electron accumulation leads to uniform spreading of signal over all the n+ strips. Hence inter-strip isolation becomes one of the major technological challenges. One of the attractive methods to achieve the inter-strip isolation is the use of uniform p-type implant on the silicon surface (p-spray). Another alternative is the use of floating p-type implants that surround the n-strips (p-stop). However, the high electric fields at the edge of the p-spray/p-stop have been shown to induce pre-breakdown micro-discharge. An optimization of the implant dose profile of the p-spray and p-stop is required to achieve good electrical isolation while ensuring satisfactory breakdown performance of the Si sensors. In the present work, we report the preliminary results of simulation study performed on the $n^+n^-$ Si sensors, equipped with p-spray and p-stops, using SILVACO tools.

physics.ins-det↗

arXiv2018

A simulation study to distinguish prompt photon from $π^0$ and beam halo in a granular calorimeter using deep networks

In a hadron collider environment identification of prompt photons originating in a hard partonic scattering process and rejection of non-prompt photons coming from hadronic jets or from beam related sources, is the first step for study of processes with photons in final state. Photons coming from decay of $π_0$'s produced inside a hadronic jet and photons produced in catastrophic bremsstrahlung by beam halo muons are two major sources of non-prompt photons. In this paper the potential of deep learning methods for separating the prompt photons from beam halo and $π^0$'s in the electromagnetic calorimeter of a collider detector is investigated, using an approximate description of the CMS detector. It is shown that, using only calorimetric information as images with a Convolutional Neural Network, beam halo (and $π^{0}$) can be separated from photon with 99.96\% (97.7\%) background rejection for 99.00\% (90.0\%) signal efficiency which is much better than traditionally employed variables.

physics.ins-det↗

arXiv2017

Simulation study of energy resolution, position resolution and $π^0$-$γ$ separation of a sampling electromagnetic calorimeter at high energies

A simulation study of energy resolution, position resolution, and $π^0$-$γ$ separation using multivariate methods of a sampling calorimeter is presented. As a realistic example, the geometry of the calorimeter is taken from the design geometry of the Shashlik calorimeter which was considered as a candidate for CMS endcap for the phase II of LHC running. The methods proposed in this paper can be easily adapted to various geometrical layouts of a sampling calorimeter. Energy resolution is studied for different layouts and different absorber-scintillator combinations of the Shashlik detector. It is shown that a boosted decision tree using fine grained information of the calorimeter can perform three times better than a cut-based method for separation of $π^0$ from $γ$ over a large energy range of 20 GeV-200 GeV.

physics.ins-det↗

arXiv2015

Probing the light radion through diphotons at the Large Hadron Collider

A radion in a scenario with a warped extra dimension can be lighter than the Higgs boson, even if the Kaluza-Klein excitation modes of the graviton turn out to be in the multi-TeV region. The discovery of such a light radion would be gateway to new physics. We show how the two-photon mode of decay can enable us to probe a radion in the mass range 60 - 110 GeV. We take into account the diphoton background, including fragmentation effects, and include cuts designed to suppress the background to the maximum possible extent. Our conclusion is that, with an integrated luminosity of 3000 $\rm fb^{-1}$ or less, the next run of the Large Hadron Collider should be able to detect a radion in this mass range, with a significance of 5 standard deviations or more.

hep-ph↗

arXiv2009

Quark Excitations Through the Prism of Direct Photon Plus Jet at the LHC

The quest to know the structure of matter has resulted in various theoretical speculations wherein additional colored fermions are postulated. Arising either as Kaluza-Klein excitations of ordinary quarks, or as excited states in scenarios wherein the quarks themselves are composites, or even in theories with extended gauge symmetry, the presence of such fermions ($q^*$) can potentially be manifested in $γ+ jet$ final states at the LHC. Using unitarized amplitudes and the CMS setup, we demonstrate that in the initial phase of LHC operation (with an integrated luminosity of $200 \pb^{-1}$) one can discover such states for a mass upto 2.0 TeV. The discovery of a $q^*$ with a mass as large as $\sim$5 TeV can be acheived for an integrated luminosity of $\sim 140 \fb^{-1}$. We also comment on the feasibility of mass determination.

hep-ph↗

arXiv2007

Search for Excited Quarks in $q\bar{q} \to γγ$ at the LHC

If quarks are composite particles, then excited states are expected to play a rôle in the Large Hadron Collider phenomena. Concentrating on virtual effects, and using a large part of the CMS detection criteria, we present here a realistic examination of their effect in diphoton production at the LHC. For various luminosities, we present the 99 % confidence limit (CL) achievable in $Λ-M_{q*}$ parameter space where $Λ$ is the compositeness scale and M_{q^*} the mass of the state. For a q^* of mass 0.5 TeV, $Λ\leq 1.55 (2.95)$ can be excluded at 99% CL with 30 (200)${\rm fb}^{-1}$ integrated luminosity.

hep-ph↗