arXiv subjects

Sanmay Ganguly

Publications and source records attributed to Sanmay Ganguly.

At least 19 recordsLinked to original sources

arXiv2026

Probing SMEFT Operators through $t\bar{t}t\bar{t}$ Production with Hyper-Graph Neural Networks at the LHC

We present a phenomenological study of $t\bar{t}t\bar{t}$ production in proton-proton collisions at $\sqrt{s} = 13$~TeV, using a Hyper-Graph Neural Network (H-GNN) to discriminate multilepton signal events from the dominant SM backgrounds, namely $t\bar{t}W$, $t\bar{t}Z$, $t\bar{t}H$, $t\bar{t}VV$, single-top associated production, and diboson and triboson processes. In the H-GNN architecture each event is represented as a hypergraph whose nodes correspond to reconstructed jets and leptons and whose hyperedges encode higher-order correlations among arbitrary subsets of these objects, allowing the network to learn the many-body kinematic structures that characterize the $t\bar{t}t\bar{t}$ final state. Combining same-sign di-lepton, tri-lepton, and four-lepton channels following a CMS-like event selection, the H-GNN attains an area under the ROC curve of $0.951$ for the $t\bar{t}t\bar{t}$ signal and yields a statistical significance of $Z = 9.11$ at an integrated luminosity of $\mathcal{L} = 140~\mathrm{fb}^{-1}$, to be compared with $Z = 8.62$ for a SPANet baseline, $Z = 7.37$ for a Particle Transformer baseline, and $Z = 5.13$ obtained by the ATLAS analysis, evaluated under identical event selection. We exploit the improved signal extraction to derive one- and two-parameter $95\%$ confidence level limits on the Wilson coefficients of the dimension-six operators $\mathcal{O}_{Φu}$, $\mathcal{O}^{(1)}_{tt}$, $\mathcal{O}^{(1)}_{qq}$, $\mathcal{O}^{(1)}_{qt}$, and $\mathcal{O}^{(8)}_{qt}$, and we project the expected sensitivity at the HL-LHC integrated luminosities of $1000~\mathrm{fb}^{-1}$ and $3000~\mathrm{fb}^{-1}$ with $50\%$ uncertainty on the background estimation.

hep-ph↗

arXiv2026

Dissecting Jet-Tagger Through Mechanistic Interpretability

Mechanistic interpretability seeks to reverse engineer a trained neural network by identifying the minimal subset of internal components. We perform a mechanistic interpretability analysis of the Particle Transformer architecture, trained on the Top Quark Tagging reference dataset, with the goal of identifying the computational circuit responsible for jet classification and characterizing the physical content of its internal representations. Combining zero ablation, path patching with two complementary on-manifold corruption strategies and linear probing of the residual stream, we identify a sparse six-head circuit that recovers the great majority of the full model performance while admitting a clean source-relay-readout interpretation. In this circuit, a single early layer head serves as the primary causal source, a cluster of middle-layer heads acts as relays selectively attending to hard pairwise substructure and a single late-layer head reads out the aggregated signal. Linear probes show that the residual stream is preferentially aligned with the energy correlator basis over the $N$-subjettiness basis. Within the energy correlator basis, the model preferentially encodes 2-prong substructure observables over the 3-prong observables. A per-layer trained probe further reveals that the apparent single step commitment of the model to a classification decision in the first class attention block is in fact a basis rotation, with the discriminating signal already saturating in the particle attention stack. These results demonstrate that mechanistic interpretability methods developed for natural language models can be used for jet physics classifiers and indicate that gradient descent may rediscover physically meaningful aspects of jet tagging without supervision.

hep-ph↗

arXiv2026

BRICKS: Compositional Neural Markov Kernels for Zero-Shot Radiation-Matter Simulation

We introduce a new strategy for compositional neural surrogates for radiation-matter interactions, a key task spanning domains from particle physics through nuclear and space engineering to medical physics. Exploiting the locality and the Markov nature of particle interactions, we create a \emph{next-particle prediction} kernel using hybrid discrete-continuous transformer models based on Riemannian Flow Matching on product manifolds. The model generates variable-sized typed sets of particles and radiation side effects that are the result of the interaction of an incident particle with a material volume. The resulting kernel can be composed to simulate unseen large-scale material distributions in a zero-shot manner. Unlike mechanistic simulators, our model is designed to be differentiable, provides tractable likelihoods for future downstream applications. A significant computational speed-up on GPU compared to CPU-bound mechanistic simulation is observed for single-kernel execution. We evaluate the model at the kernel level and demonstrate predictive stability over multi-round autoregressive rollouts. We additionally release a novel 20M-event radiation-matter interaction dataset for further research.

cs.LG↗

arXiv2026

Explainable AI for Jet Tagging: A Comparative Study of GNNExplainer, GNNShap, and GradCAM for Jet Tagging in the Lund Jet Plane

Graph neural networks such as ParticleNet and transformer based networks on point clouds such as ParticleTransformer achieve state-of-the-art performance on jet tagging benchmarks at the Large Hadron Collider, yet the physical reasoning behind their predictions remains opaque. We present different methods, i.e. perturbation-based (GNNExplainer), Shapley-value-based (GNNShap), and gradient-based (GRADCam); adapted to operate on LundNet's Lund-plane graph representation. Leveraging the fact that each node in the Lund plane corresponds to a physically meaningful parton splitting, we construct Monte Carlo truth explanation masks and introduce a physics-informed evaluation framework that goes beyond standard fidelity metrics. We perform the analysis in three transverse-momentum bins ($\mathrm{p_T} \in [500,700]$, $[800,1000]$, and the inclusive region $[500,1000]$ GeV), revealing how explanation quality and focus shift between non-perturbative and perturbative regimes. We further quantify the correlation between explainer-assigned node importance and classical jet substructure observables -- $N$-subjettiness ratios $τ_{21}$ and $τ_{32}$ and the energy correlation functions -- establishing the degree to which the model has learned known QCD features. We find that overall the weight assigned by explainability methods has a correlation with analytic observables, with expected shift across different phase space regimes, indicating that a trained neural network indeed learns some aspects of jet-substructure moments. Our open-source implementation enables reproducible explainability studies for graph-based jet taggers.

hep-ph↗

arXiv2025

HHH Whitepaper

We here report on the progress of the HHH Workshop, that took place in Dubrovnik in July 2023. After the discovery of a particle that complies with the properties of the Higgs boson of the Standard Model, all Standard Model (SM) parameters are in principle determined. However, in order to verify or falsify the model, the full form of the potential has to be determined. This includes the measurement of the triple and quartic scalar couplings. We here report on ongoing progress of measurements for multi-scalar final states, with an emphasis on three SM-like scalar bosons at 125 GeV, but also mentioning other options. We discuss both experimental progress and challenges as well as theoretical studies and models that can enhance such rates with respect to the SM predictions

hep-ph↗

arXiv2024

Velocity recostruction with graph neural networks

In this work, we seek to improve the velocity reconstruction of clusters by using Graph Neural Networks -- a type of deep neural network designed to analyze sparse, unstructured data. In comparison to the Convolutional Neural Network (CNN) which is built for structured data such as regular grids, GNN is particularly suitable for analyzing galaxy catalogs. In our GNNs, galaxies as represented as nodes that are connected with edges. The galaxy positions and properties -- stellar mass, star formation rate, and total number of galaxies within 100~\mpc -- are combined to predict the line-of-sight velocity of the clusters. To train our networks, we use mock SDSS galaxies and clusters constructed from the Magneticum hydrodynamic simulations. Our GNNs reach a precision in reconstructed line-of-sight velocity of $Δv$=163 km/s, outperforming by $\approx$10\% the perturbation theory~($Δv$=181 km/s) or the CNN~($Δv$=179 km/s). The stellar mass provides additional information, improving the precision by $\approx$6\% beyond the position-only GNN, while other properties add little information. Our GNNs remain capable of reconstructing the velocity field when redshift-space distortion is included, with $Δv$=210 km/s which is again 10\% better than CNN with RSD. Finally, we find that even with an impressive, nearly 70\% increase in galaxy number density from SDSS to DESI, our GNNs only show an underwhelming 2\% improvement, in line with previous works using other methods. Our work demonstrates that, while the efficiency in velocity reconstruction may have plateaued already at SDSS number density, further improvements are still hopeful with new reconstruction models such as the GNNs studied here.

astro-ph.CO↗

arXiv2023

Set-Conditional Set Generation for Particle Physics

The simulation of particle physics data is a fundamental but computationally intensive ingredient for physics analysis at the Large Hadron Collider, where observational set-valued data is generated conditional on a set of incoming particles. To accelerate this task, we present a novel generative model based on a graph neural network and slot-attention components, which exceeds the performance of pre-existing baselines.

hep-ex↗

arXiv2023

Reconstructing particles in jets using set transformer and hypergraph prediction networks

The task of reconstructing particles from low-level detector response data to predict the set of final state particles in collision events represents a set-to-set prediction task requiring the use of multiple features and their correlations in the input data. We deploy three separate set-to-set neural network architectures to reconstruct particles in events containing a single jet in a fully-simulated calorimeter. Performance is evaluated in terms of particle reconstruction quality, properties regression, and jet-level metrics. The results demonstrate that such a high dimensional end-to-end approach succeeds in surpassing basic parametric approaches in disentangling individual neutral particles inside of jets and optimizing the use of complementary detector information. In particular, the performance comparison favors a novel architecture based on learning hypergraph structure, HGPflow, which benefits from a physically-interpretable approach to particle reconstruction.

hep-ex↗

arXiv2023

Configurable calorimeter simulation for AI applications

A configurable calorimeter simulation for AI (COCOA) applications is presented, based on the Geant4 toolkit and interfaced with the Pythia event generator. This open-source project is aimed to support the development of machine learning algorithms in high energy physics that rely on realistic particle shower descriptions, such as reconstruction, fast simulation, and low-level analysis. Specifications such as the granularity and material of its nearly hermetic geometry are user-configurable. The tool is supplemented with simple event processing including topological clustering, jet algorithms, and a nearest-neighbors graph construction. Formatting is also provided to visualise events using the Phoenix event display software.

hep-ex↗

arXiv2022

Machine Learning and LHC Event Generation

First-principle simulations are at the heart of the high-energy physics research program. They link the vast data output of multi-purpose detectors with fundamental theory predictions and interpretation. This review illustrates a wide range of applications of modern machine learning to event generation and simulation-based inference, including conceptional developments driven by the specific requirements of particle physics. New ideas and tools developed at the interface of particle physics and machine learning will improve the speed and precision of forward simulations, handle the complexity of collision data, and enhance inference as an inverse simulation problem.

hep-ph↗

arXiv2022

Improving Di-Higgs Sensitivity at Future Colliders in Hadronic Final States with Machine Learning

One of the central goals of the physics program at the future colliders is to elucidate the origin of electroweak symmetry breaking, including precision measurements of the Higgs sector. This includes a detailed study of Higgs boson (H) pair production, which can reveal the H self-coupling. Since the discovery of the Higgs boson, a large campaign of measurements of the properties of the Higgs boson has begun and many new ideas have emerged during the completion of this program. One such idea is the use of highly boosted and merged hadronic decays of the Higgs boson ($\mathrm{H}\to\mathrm{b}\bar{\mathrm{b}}$, $\mathrm{H}\to\mathrm{W}\mathrm{W}\to\mathrm{q}\bar{\mathrm{q}}\mathrm{q}\bar{\mathrm{q}}$) with machine learning methods to improve the signal-to-background discrimination. In this white paper, we champion the use of these modes to boost the sensitivity of future collider physics programs to Higgs boson pair production, the Higgs self-coupling, and Higgs-vector boson couplings. We demonstrate the potential improvement possible at the Future Circular Collider in hadron mode, especially with the use of graph neural networks.

hep-ph↗

arXiv2022

Graph Neural Networks in Particle Physics: Implementations, Innovations, and Challenges

Many physical systems can be best understood as sets of discrete data with associated relationships. Where previously these sets of data have been formulated as series or image data to match the available machine learning architectures, with the advent of graph neural networks (GNNs), these systems can be learned natively as graphs. This allows a wide variety of high- and low-level physical features to be attached to measurements and, by the same token, a wide variety of HEP tasks to be accomplished by the same GNN architectures. GNNs have found powerful use-cases in reconstruction, tagging, generation and end-to-end analysis. With the wide-spread adoption of GNNs in industry, the HEP community is well-placed to benefit from rapid improvements in GNN latency and memory usage. However, industry use-cases are not perfectly aligned with HEP and much work needs to be done to best match unique GNN capabilities to unique HEP obstacles. We present here a range of these capabilities, predictions of which are currently being well-adopted in HEP communities, and which are still immature. We hope to capture the landscape of graph techniques in machine learning as well as point out the most significant gaps that are inhibiting potentially large leaps in research.

hep-ex↗

arXiv2022

Symmetry Group Equivariant Architectures for Physics

Physical theories grounded in mathematical symmetries are an essential component of our understanding of a wide range of properties of the universe. Similarly, in the domain of machine learning, an awareness of symmetries such as rotation or permutation invariance has driven impressive performance breakthroughs in computer vision, natural language processing, and other important applications. In this report, we argue that both the physics community and the broader machine learning community have much to understand and potentially to gain from a deeper investment in research concerning symmetry group equivariant machine learning architectures. For some applications, the introduction of symmetries into the fundamental structural design can yield models that are more economical (i.e. contain fewer, but more expressive, learned parameters), interpretable (i.e. more explainable or directly mappable to physical quantities), and/or trainable (i.e. more efficient in both data and computational requirements). We discuss various figures of merit for evaluating these models as well as some potential benefits and limitations of these methods for a variety of physics applications. Research and investment into these approaches will lay the foundation for future architectures that are potentially more robust under new computational paradigms and will provide a richer description of the physical systems to which they are applied.

cs.LG↗

arXiv2021

Secondary Vertex Finding in Jets with Neural Networks

Jet classification is an important ingredient in measurements and searches for new physics at particle coliders, and secondary vertex reconstruction is a key intermediate step in building powerful jet classifiers. We use a neural network to perform vertex finding inside jets in order to improve the classification performance, with a focus on separation of bottom vs. charm flavor tagging. We implement a novel, universal set-to-graph model, which takes into account information from all tracks in a jet to determine if pairs of tracks originated from a common vertex. We explore different performance metrics and find our method to outperform traditional approaches in accurate secondary vertex reconstruction. We also find that improved vertex finding leads to a significant improvement in jet classification performance.

hep-ex↗

arXiv2021

Towards a Computer Vision Particle Flow

In High Energy Physics experiments Particle Flow (PFlow) algorithms are designed to provide an optimal reconstruction of the nature and kinematic properties of the particles produced within the detector acceptance during collisions. At the heart of PFlow algorithms is the ability to distinguish the calorimeter energy deposits of neutral particles from those of charged particles, using the complementary measurements of charged particle tracking devices, to provide a superior measurement of the particle content and kinematics. In this paper, a computer vision approach to this fundamental aspect of PFlow algorithms, based on calorimeter images, is proposed. A comparative study of the state of the art deep learning techniques is performed. A significantly improved reconstruction of the neutral particle calorimeter energy deposits is obtained in a context of large overlaps with the deposits from charged particles. Calorimeter images with augmented finer granularity are also obtained using super-resolution techniques.

physics.data-an↗

arXiv2020

Light quark Yukawas in triboson final states

Triple heavy vector boson production, $p p \to VVV$ $(V = W, Z)$, has recently been observed for the first time. We propose that precision measurements of this process provide an excellent probe of the first generation light quark Yukawa couplings. Modified quark interactions with the off-shell Higgs in this process lead to a rapid growth of the partonic cross sections with energy, which manifests in an enhanced $p_T$ distribution of the final state leptons and quarks. We quantify this effect and estimate the present and future 2$σ$ sensitivity to the up, down, and strange Yukawas. In particular, we find that HL-LHC can reach $\mathcal{O}(400)$ sensitivity to the down Yukawa relative to the Standard Model value, improving the current sensitivity in this process by a factor of $10$, and which can be further improved to $\mathcal{O}(30)$ at FCC-hh. This is competitive with and complementary to constraints from global fits and other on-shell probes of the first generation Yukawas. The triboson sensitivity at HL-LHC corresponds to probing dimension-6 SMEFT operators suppressed by an $\mathcal{O}(1)$ TeV scale, similarly to other LHC Higgs probes.

hep-ph↗

arXiv2013

Jet Production Measurements at CMS

Jet production cross-section measurements are presented. The measurements are done with the data from Large Hadron Collider (LHC) proton-proton collisions, collected with the Compact Muon Solenoid (CMS) detector. The inclusive jet production measurements are carried out with data collected $\rm \sqrt{s} ~= ~7 ~TeV$ and $\rm 8~TeV$ with total integrated luminosity ($\mathcal{L}_{int}$) $\rm 5.0~ fb^{-1}$ and $\rm 10.71~ fb^{-1}$ respectively. The dijet production measurements are carried out with the $\rm \sqrt{s}~ =~ 7 ~TeV$ dataset. Jets are reconstructed with the anti-$k_T$ clustering algorithm with size parameter $R=0.7$. The measured cross sections are corrected for detector effects and compared to perturbative QCD predictions at NLO, corrected for NP factors, using various sets of PDF. The inclusive jet cross-section ratio of the jets reconstructed with the anti-$k_T$ (AK) algorithm and two radius parameter $\rm R~=~0.5$ and $\rm R~=~0.7$ are also presented. The data used is $\rm \sqrt{s}~ =~ 7 ~TeV$ CMS data corresponding to $\rm \mathcal{L}_{int}~=~5.0 ~fb^{-1}$. Significant discrepancies are found comparing the data to leading order calculations and to fixed order calculations at NLO, corrected for NP effects, whereas simulations with NLO matrix elements matched to the parton showers describe the data quite well. A study of color coherence effects in pp collisions has been performed with the data collected at $\rm \sqrt{s}~ =~ 7~TeV$ and $\rm\mathcal{L}_{int}~=~ 36~pb^{-1}$. The measurement of the azimuthal angular correlation between the second and third jets is compared to the predictions of Monte Carlo models with different implementations of color coherence effects.

hep-ex↗

arXiv2013

Jet Cross-Section Measurements In CMS

The Large Hadron Collider (LHC) experiment has successfully completed data taking at center of mass (COM) energy 7 TeV in 2011 and very recently for 8 TeV. Measurement of cross sections predicted by the standard model were the main tasks in the beginning. The inclusive jet cross section and dijet mass measurement is already done at 7 TeV energy by Compact Muon Solenoid (CMS) detector with integrated luminosity 5 fb$^{-1}$. In these measurement one needs to understand and measure precisely the kinematic properties of jets which involve many theoretical and experimental issues. The goal of this article is to discuss all these issues including jet measurements in CMS and subsequently review the inclusive jet cross section and dijet mass measurement in CMS at 7 TeV with integrated luminosity 5 fb$^{-1}$. The measurements, after unfolding the data, are also compared with the next leading order (NLO) theory predictions, corrected for the non-perturbative (NP) effects, for five different sets of parton distribution functions (PDF). It is observed that the measurements, for both cases, agree with the theory prediction within $\sim$8-10% depending on transverse momentum ($p_T$) and dijet invariant mass ($M_{jj}$) of jets.

hep-ph↗