arXiv subjects

Alejandro Szynkman

Publications and source records attributed to Alejandro Szynkman.

At least 19 recordsLinked to original sources

arXiv2026

Two-Stage Machine Learning Strategy for Scalar and Vector Leptoquark Discrimination at the LHC

We present a machine learning framework for the characterization of leptoquark (LQ) signals at the Large Hadron Collider, focusing on the discrimination between scalar (SLQ) and vector (VLQ) hypotheses. The method is based on a two-stage inference pipeline that combines a classifier trained to separate Standard Model backgrounds from a mixed LQ signal with a second classifier designed to distinguish between SLQ and VLQ scenarios, using the signal yield inferred from the first-stage classifier to guide the corresponding scalar and vector mass hypotheses. A test statistic is constructed from the classifier outputs and interpreted using reference probability density functions. The approach is applied to realistic LHC final states with hadronically decaying tau leptons, multiple jets, and missing transverse momentum, and its performance is assessed using simulated pseudo-experiments. We show that the proposed strategy provides a robust and statistically consistent procedure to discriminate between SLQ and VLQ signals across a wide range of masses and signal stregths. We find that the spin identification power closely follows the discovery potential, demonstrating that determining the spin nature of a newly discovered LQ does not require substantially larger datasets than those needed for discovery itself.

hep-ph↗

arXiv2025

Exploring enhanced non-resonant di-Higgs production at the HL-LHC with neural networks

We investigate di-Higgs production in the $b\bar{b}γγ$ final state at the LHC, focusing on scenarios where the gluon fusion process is enhanced by new colored scalars, which could be identified as squarks or leptoquarks. We consider two benchmarks characterized by the mass of the lightest colored scalar, BM$_{\mathrm{L}}$ and BM$_{\mathrm{H}}$, corresponding to 464 GeV and 621 GeV, respectively. Using Monte Carlo simulations for both the signal and the dominant backgrounds, we perform a discovery analysis with deep neural networks, exploring various architectures and input variables. Our results show that the discrimination power is maximized by employing two dedicated classifiers, one trained against QCD backgrounds and another against backgrounds involving single-Higgs processes. Furthermore, we demonstrate that including high-level features -- such as the invariant masses $m_{γγ}$, $m_{bb}$, and $m_{hh}$, as well as the transverse momenta and angular separations of the photon and $b$-jet pairs -- significantly improves the performance compared to using only low-level features as the invariant mass and momenta of the final particles. For the latter case, we find that architectures processing photon and $b$-jet variables separately can enhance the significance for BM$_{\mathrm{H}}$. Projecting for an integrated luminosity of 3 ab$^{-1}$, we obtain a significance of 7.3 for BM$_{\mathrm{L}}$, while it drops to 3.1 for BM$_{\mathrm{H}}$. In the particular case of BM$_{\mathrm{L}}$, discovery level significance can be reached at 1.7 ab$^{-1}$.

hep-ph↗

arXiv2025

Di-Higgs to 4b with Bayesian inference: improving simulation estimates

Measuring di-Higgs production in the four-bottom channel is challenged by overwhelming QCD backgrounds and imperfect simulations. We develop a Bayesian mixture model that simultaneously infers signal and background fractions and their individual shapes directly in the signal region. The likelihood is a nuanced combination of a one-dimensional kinematic discriminator and per-jet flavour scores; with their correlations incorporated via kinematic bins. Monte Carlo informs weak Dirichlet priors, while the posterior adjusts to the interplay of the model, priors and observed data. Using pseudo-data simulated with standard tools and with controlled mismatches, we show that the method corrects biased priors, delivers calibrated 68-95% credible intervals for the signal count, and improves dataset-level ROC/AUC relative to simple cut-and-count baselines. This study highlights how Bayesian inference can harvest information present in the signal region and self-calibrate model parameters, providing a robust route to increased sensitivity in di-Higgs searches.

hep-ph↗

arXiv2025

Inferring correlated distributions: boosted top jets

Improving the understanding of signal and background distributions in signal-region is a valuable key to enhance any analysis in collider physics. This is usually a difficult task because -- among others -- signal and backgrounds are hard to discriminate in signal-region, simulations may reach a limit of reliability if they need to model non-perturbative QCD, and distributions are multi-dimensional and many times may be correlated within each class. Bayesian density estimation is a technique that leverages prior knowledge and data correlations to effectively extract information from data in signal-region. In this work we extend previous works on data-driven mixture models for meaningful unsupervised signal extraction in collider physics to incorporate correlations between features. Using a standard dataset of top and QCD jets, we show how simulators, despite having an expected bias, can be used to inject sufficient inductive nuance into an inference model in terms of priors to then be corrected by data and estimate the true correlated distributions between features within each class. We compare the model with and without correlations to show how the signal extraction is sensitive to their inclusion and we quantify the improvement due to the inclusion of correlations using both supervised and unsupervised metrics.

hep-ph↗

arXiv2025

Probing triple-gauge couplings in anomalous gauge theories at hadron and lepton colliders

Gauge anomalous quantum field theories are inconsistent as full UV theories since they lead to the breaking of Lorentz invariance or Unitarity, as well as non-renormalizability. It is well known, however, that they can be interpreted as effective field theories (EFT) with a cut-off. The latter cannot be made arbitrarily large and it is related to the energy scale at which additional fermions with suitable gauge charges enter, rendering the full model anomaly-free. A nondecoupling effect that remains in the EFT is the appearance of anomalous loop-induced triple-gauge couplings, encapsulating information from the full UV theory. In this work we take as an example an Abelian gauge symmetry $U(1)'_μ$ under which $2^{nd}$-generation leptons are axially charged, leading to an EFT that consists of the Standard Model (SM) with an additional massive $Z'$ gauge boson. As a consequence, there are triple gauge couplings involving the $Z'$ and Electroweak SM gauge bosons via mixed gauge anomalies. We study the possibility of probing these loop suppressed anomalous couplings at hadron and lepton colliders, with $Z'$-lepton couplings allowed by current experimental bounds, finding that due to the large SM backgrounds and small signal, the HL-LHC is incapable of this task. The 100 TeV $pp$ collider at $\mathcal{L}=20~\mathrm {ab}^{-1}$ on the other hand could probe anomalous couplings for $m_{Z'}\in[150,800]~\mathrm{GeV}$ and obtain discovery significances for $m_{Z'}\in[230,330]~\mathrm{GeV}$. Lepton colliders are also well suited for probing these anomalous couplings. In particular we show that a muon collider running at the $Z'$-resonance and an electron-positron collider such as CLIC with $\sqrt{s}=3~{\rm TeV}$ can be complimentary in probing the anomalous couplings for $m_{Z'}\in[100,700]~{\rm GeV}$, with CLIC sensitive to discovery for $m_{Z'}\in[125,225]~{\rm GeV}$.

hep-ph↗

arXiv2024

Tripartite entanglement from experimental data: $B^0\to K^{*0}μ^+μ^-$ as a case study

We develop an angular analysis based on the reconstruction of the helicity amplitudes from dedicated experimental data corresponding to the tripartite state composed by one qutrit and two qubits, which arises in the three-body decays of a spin zero particle into one vector and a fermion pair. Starting from the associated spin density matrix of the final state, entanglement quantifiers were investigated and the corresponding significances were determined up to second order in the error propagation of the uncertainties of the angular measurements. As an application of our analysis, we performed a full quantum tomography of the final state in the $B^0\to K^{*0}μ^+μ^-$ decays using data recorded by LHCb collaboration. We found the presence of genuine quantum entanglement of the final state and also in both kaon-muon and di-muon subsystems. In recent years, $B$ meson decays received significant attention from both experimental and theoretical sides, and the proposed observables provide novel perspectives for studying them. Furthermore, this analysis could be also applied to other several processes if the complete experimental data were available for the helicity amplitudes reconstruction.

hep-ph↗

arXiv2024

Improvement and generalization of ABCD method with Bayesian inference

To find New Physics or to refine our knowledge of the Standard Model at the LHC is an enterprise that involves many factors. We focus on taking advantage of available information and pour our effort in re-thinking the usual data-driven ABCD method to improve it and to generalize it using Bayesian Machine Learning tools. We propose that a dataset consisting of a signal and many backgrounds is well described through a mixture model. Signal, backgrounds and their relative fractions in the sample can be well extracted by exploiting the prior knowledge and the dependence between the different observables at the event-by-event level with Bayesian tools. We show how, in contrast to the ABCD method, one can take advantage of understanding some properties of the different backgrounds and of having more than two independent observables to measure in each event. In addition, instead of regions defined through hard cuts, the Bayesian framework uses the information of continuous distribution to obtain soft-assignments of the events which are statistically more robust. To compare both methods we use a toy problem inspired by $pp\to hh\to b\bar b b \bar b$, selecting a reduced and simplified number of processes and analysing the flavor of the four jets and the invariant mass of the jet-pairs, modeled with simplified distributions. Taking advantage of all this information, and starting from a combination of biased and agnostic priors, leads us to a very good posterior once we use the Bayesian framework to exploit the data and the mutual information of the observables at the event-by-event level. We show how, in this simplified model, the Bayesian framework outperforms the ABCD method sensitivity in obtaining the signal fraction in scenarios with $1\%$ and $0.5\%$ true signal fractions in the dataset. We also show that the method is robust against the absence of signal.

hep-ph↗

arXiv2024

Double Higgs production at the HL-LHC: probing a loop-enhanced model with kinematical distributions

We study di-Higgs production via gluon fusion at the high luminosity LHC in the presence of new physics, focusing on the $b\bar bγγ$ final states. Taking a minimal set of three scalar leptoquarks (LQs) with cubic and quartic interactions with the Higgs and choosing four benchmark points with a light LQ, we perform a detailed analysis of differential distributions of the di-Higgs production cross section, studying the imprints of the new physics states running in the loops. Simulating the signal and main backgrounds, we study the influence of the new physics in differential distributions such as the invariant mass of the subsystems of final particles, the transverse momentum, and angular variables, finding in particular a resonance peak associated with the light LQ. It turns out that the angular separation of the photons, which is correlated with the resonance LQ peak, is a very sensitive observable that helps in discriminating the new physics signal from the Standard Model background. We find that for two of our benchmarks discovery could be reached with 3 ab$^{-1}$, whereas exclusion limits at 95% C.L. could be claimed with 0.60-0.75 ab$^{-1}$. For the other two benchmarks that have heavier LQ masses significances of order 2$σ$ are possible for 3 ab$^{-1}$. A similar analysis could be applied to other loop-enhanced models.

hep-ph↗

arXiv2024

LHC Study of Third-Generation Scalar Leptoquarks with Machine-Learned Likelihoods

We study the impact of machine-learning algorithms on LHC searches for leptoquarks in final states with hadronically decaying tau leptons, multiple $b$-jets, and large missing transverse momentum. Pair production of scalar leptoquarks with decays only into third-generation leptons and quarks is assumed. Thanks to the use of supervised learning tools with unbinned methods to handle the high-dimensional final states, we consider simple selection cuts which would possibly translate into an improvement in the exclusion limits at the 95$\%$ confidence level for leptoquark masses with different values of their branching fraction into charged leptons. In particular, for intermediate branching fractions, we expect that the exclusion limits for leptoquark masses extend to $\sim$1.3 TeV. As a novelty in the implemented unbinned analysis, we include a simplified estimation of some systematic uncertainties with the aim of studying their possible impact on the stability of the results. Finally, we also present the projected sensitivity within this framework at 14 TeV for 300 and 3000 fb$^{-1}$ that extends the upper limits to $\sim$1.6 and $\sim$1.8 TeV, respectively.

hep-ph↗

arXiv2023

Exploring unsupervised top tagging using Bayesian inference

Recognizing hadronically decaying top-quark jets in a sample of jets, or even its total fraction in the sample, is an important step in many LHC searches for Standard Model and Beyond Standard Model physics as well. Although there exists outstanding top-tagger algorithms, their construction and their expected performance rely on Montecarlo simulations, which may induce potential biases. For these reasons we develop two simple unsupervised top-tagger algorithms based on performing Bayesian inference on a mixture model. In one of them we use as the observed variable a new geometrically-based observable $\tilde{A}_{3}$, and in the other we consider the more traditional $τ_{3}/τ_{2}$ $N$-subjettiness ratio, which yields a better performance. As expected, we find that the unsupervised tagger performance is below existing supervised taggers, reaching expected Area Under Curve AUC $\sim 0.80-0.81$ and accuracies of about 69% $-$ 75% in a full range of sample purity. However, these performances are more robust to possible biases in the Montecarlo that their supervised counterparts. Our findings are a step towards exploring and considering simpler and unbiased taggers.

hep-ph↗

arXiv2023

Disentangling the Seesaw in the Left-Right Model -- An Algorithm for the General Case

Senjanovic and Tello have analyzed how one could determine the neutrino Dirac mass matrix in the minimal left-right model, assuming that the mass matrices for the light and heavy neutrinos could be taken as inputs. They have provided an analytical solution for the Dirac mass matrix in the case that the left-right symmetry is implemented via a generalized parity symmetry and that this symmetry remains unbroken in the Dirac Yukawa sector. We extend the work of Senjanovic and Tello to the case in which the generalized parity symmetry is broken in the Dirac Yukawa sector. In this case the elegant method outlined by Senjanovic and Tello breaks down and we need to adopt a numerical approach. Several iterative approaches are described; these are found to work in some cases but to be highly unstable in others. A stable, prescriptive numerical algorithm is described that works in all but a vanishingly small number of cases. We apply this algorithm to numerical data sets that are consistent with current experimental constraints on neutrino masses and mixings. We also provide some additional context and supporting explanations for the case in which the parity symmetry is unbroken.

hep-ph↗

arXiv2022

A method for approximating optimal statistical significances with machine-learned likelihoods

Machine-learning techniques have become fundamental in high-energy physics and, for new physics searches, it is crucial to know their performance in terms of experimental sensitivity, understood as the statistical significance of the signal-plus-background hypothesis over the background-only one. We present here a simple method that combines the power of current machine-learning techniques to face high-dimensional data with the likelihood-based inference tests used in traditional analyses, which allows us to estimate the sensitivity for both discovery and exclusion limits through a single parameter of interest, the signal strength. Based on supervised learning techniques, it can perform well also with high-dimensional data, when traditional techniques cannot. We apply the method to a toy model first, so we can explore its potential, and then to a LHC study of new physics particles in dijet final states. Considering as the optimal statistical significance the one we would obtain if the true generative functions were known, we show that our method provides a better approximation than the usual naive counting experimental results.

hep-ph↗

arXiv2022

The Elusive Muonic WIMP

The Weakly Interacting Massive Particle (WIMP) paradigm is one of the most popular scenarios for Dark Matter (DM) theories that however is strongly constrained, in particular by direct detection experiments. We stick with the WIMP hypothesis and consider a Dirac fermion candidate for DM that interacts with the Standard Model (SM) via a spin-1 $Z'$, arising from the spontaneous breaking of an Abelian $U(1)'_μ$ gauge symmetry, under which only second generation leptons and the DM are appropriately charged. Due to the charge assignment, the model is gauge anomalous and can only be interpreted as an effective field theory (EFT) at low energy. The $Z'$ couples at tree level only to the vector DM current, to the axial muon current and to left-handed muonic neutrinos, so the WIMP-nucleon cross section is beyond the experimental reach of spin-independent (SI) direct detection searches. We study the current bounds on this model coming from direct and indirect detection of DM, collider searches, contributions to $(g-2)_μ$ and to neutrino trident production. We find that large regions of the parameter space remains to be explored. In the context of LHC searches, we study the impact of a muon-exclusive signal region for the $3μ$ + ${E}^{\rm miss}_T$ channel with an invariant mass window around $m_{Z'}$. We show that this search can significantly improve the current collider bounds. Finally, from the anomalous nature of our EFT, there remain at low energy triboson anomalous interactions between the $Z'$ and the electroweak (EW) SM gauge bosons. We explore the possibilities of probing these interactions at the LHC and at a 100 TeV proton collider finding it extremely challenging. On the other hand, for a muon collider the resonant channel $μ^{+}μ^{-}\to Z'\to ZZ$ could be discovered in the most promising scenario with luminosity of $\mathcal{O}({\rm few}\; 10)$ ${\rm fb}^{-1}$.

hep-ph↗

arXiv2021

Towards a method to anticipate dark matter signals with deep learning at the LHC

We study several simplified dark matter (DM) models and their signatures at the LHC using neural networks. We focus on the usual monojet plus missing transverse energy channel, but to train the algorithms we organize the data in 2D histograms instead of event-by-event arrays. This results in a large performance boost to distinguish between standard model (SM) only and SM plus new physics signals. We use the kinematic monojet features as input data which allow us to describe families of models with a single data sample. We found that the neural network performance does not depend on the simulated number of background events if they are presented as a function of $S/\sqrt{B}$, where $S$ and $B$ are the number of signal and background events per histogram, respectively. This provides flexibility to the method, since testing a particular model in that case only requires knowing the new physics monojet cross section. Furthermore, we also discuss the network performance under incorrect assumptions about the true DM nature. Finally, we propose multimodel classifiers to search and identify new signals in a more general way, for the next LHC run.

hep-ph↗

arXiv2021

Enhancement of the double Higgs production via leptoquarks at the LHC

Measurements of single Higgs production and its decays are in good agreement with the Standard Model. There is still room for large modifications in double Higgs production at LHC, though these effects may be correlated with large corrections to other observables, in particular single Higgs production. In this work we address the issue of enhancing double Higgs production in the presence of scalar leptoquarks while satisfying all experimental constraints. We show at leading order that including more than one species of leptoquarks, large cubic interactions with the Higgs can lead to sizable enhancement of di-Higgs production cross section at LHC, while at the same time keeping other Higgs observables and precision measurements under control. For masses above 800 GeV these corrections are in general below 30%, whereas in a viable scenario in which one of the leptoquarks can be light, specifically in the mass range $400-600$ GeV, we show that it is possible to roughly double the SM cross section for di-Higgs production, implying that possible first hints of it may be probed at the high luminosity LHC at $\mathcal{L}\sim 2$ ab$^{-1}$.

hep-ph↗

arXiv2019

Model-independent search strategy for the lepton-flavor-violating heavy Higgs boson decay to $τμ$ at the LHC

In this work we present a model-independent search strategy at the LHC for heavy Higgs bosons decaying into a tau and a muon, $H/A \rightarrow τμ$, showing a plausible tendency to improve the sensitivity obtained by the present experimental limits. This search strategy is performed for the Higgs boson mass range 1-5 TeV and uses as the most relevant kinematical variables, in order to discriminate signal against background, the transverse momenta of the muon and the tau together with the missing transverse energy. We estimate the exclusion limits at 95% C.L. and the significances for evidence and discovery at $\sqrt{s}$ = 14 TeV with $\cal{L}$ = 300 fb$^{-1}$, observing a growth in the sensitivities for high Higgs boson masses. Moreover, since the Higgs boson decay into a $τ$-lepton pair may mimic our LFV signal, we also study the impact of the ditau channel on the exclusion limits and the significances for evidence and discovery. In particular, the impact on the exclusion limits of LFV heavy Higgs boson decays is significant when the ditau rate begins to compete with the corresponding to the $H/A \rightarrow τμ$ decay.

hep-ph↗

arXiv2019

Constraining R-axion models through dijet searches at the LHC

The search at hadron colliders for new massive resonances of a few 100 GeVs that couple effectively to colored states is an extremely challenging issue, due principally to the presence of large QCD multijet backgrounds at this energy, rendering the searches at the LHC particularly difficult. Recently, it was realized that these large backgrounds could be overcome by demanding one high-$p_T$ jet from initial-state radiation (ISR) and by means of novel jet-reconstruction techniques through which the resulting hadronized products of the massive resonances are reconstructed as a fat-jet, a unique large-radius jet. The ATLAS and CMS Collaborations have recently reported searches for the experimental signature of a single fat-jet in association with an ISR jet. Models of dynamical supersymmetry breaking with an spontaneously broken R-symmetry give rise to the appearance of a pseudo-Nambu-Goldstone boson called the R-axion, which naturally tends to be light. In the parameter space regions where the anomalous R-axion coupling to gluons is boosted, these models can be tested against these new LHC dijet searches. Taking into account the CMS search, we apply the $q_μ$ statistical method to the signal events against the background-only expectation and obtain the 95\% C.L. exclusion limits on the most relevant model parameters for a particular messenger sector, namely, the R-axion mass $m_a$, the decay constant $f_a$, and the number of color messengers $N$, being these limits suitable to be applied to more general models with axion-like particles.

hep-ph↗

arXiv2018

Single-top production and rare top interactions

The study of the top quark's properties is an important part of the LHC programme. In earlier work, we have studied the rare decay $t \to b \bar b c$, using effective operators to capture the effects of physics beyond the Standard Model. However top decay is primarily sensitive to new physics in the sub-TeV energy regime. If this new physics resides at a higher energy scale, then one needs to turn to single-top production. In this paper, we use the $s$-channel and $t$-channel single-top production measurements to constrain the new physics parameter space associated with such contact interactions. We also study the net top polarization as a means to distinguish between contributions from operators involving different fermion chiralities and Lorentz structures.

hep-ph↗