Neural simulation-based inference on unbinned top-quark pair data at 13 TeV yields improved gluon PDF precision over traditional binned analyses while incorporating experimental and theoretical uncertainties.
hub Tool reference
DELPHES 3, A modular framework for fast simulation of a generic collider experiment
Tool reference. 79% of classified Pith citations use this work as a method, library, or software dependency, not as a substantive claim.
72
Pith papers citing it
2,080
external citations · Pith
Method reference
79% of classified citations
abstract
The version 3.0 of the DELPHES fast-simulation is presented. The goal of DELPHES is to allow the simulation of a multipurpose detector for phenomenological studies. The simulation includes a track propagation system embedded in a magnetic field, electromagnetic and hadron calorimeters, and a muon identification system. Physics objects that can be used for data analysis are then reconstructed from the simulated detector response. These include tracks and calorimeter deposits and high level objects such as isolated electrons, jets, taus, and missing energy. The new modular approach allows for greater flexibility in the design of the simulation and reconstruction sequence. New features such as the particle-flow reconstruction approach, crucial in the first years of the LHC, and pile-up simulation and mitigation, which is needed for the simulation of the LHC detectors in the near future, have also been implemented. The DELPHES framework is not meant to be used for advanced detector studies, for which more accurate tools are needed. Although some aspects of DELPHES are hadron collider specific, it is flexible enough to be adapted to the needs of electron-positron collider experiments.
hub tools
citation-role summary
method 15
background 3
baseline 1
citation-polarity summary
representative citing papers
RANs generalize moment unfolding to full phase-space unbinned unfolding via detector-level Wasserstein critics without requiring support overlap or multiple iterations.
PLuM multimodal transformer improves top and H->bb jet tagging by jointly processing particle constituents and Lund plane splittings, yielding 25% higher background rejection at 25% di-Higgs efficiency.
An XFEL Compton gamma-gamma collider at 125 GeV with a set transformer deep learning classifier on particle-flow point clouds can achieve high-precision Higgs measurements across hadronic, semi-leptonic, and leptonic final states including H to strange quarks.
Collider-Bench is a new benchmark showing that current LLM agents cannot reliably reproduce LHC analyses at the level of a physicist-in-the-loop.
Higher-order Fisher tensors in exponential-family coordinates of binned energy correlators are simultaneously local KL coefficients, connected cumulants, and hyperedge weights, enabling hypergraph constructions for jet substructure analysis.
Template-Adapted Mixture Model uses many biased simulations for data-driven estimates of signal and background distributions, yielding unbiased signal fraction estimates with well-calibrated uncertainties.
Neural refinement of Monte Carlo sample weights via phase-space scaling and a new resampling protocol that maintains averages and uncertainties.
MadGraph5_aMC@NLO automates tree-level, NLO, shower-matched, and merged cross-section computations for collider processes in a unified flexible framework.
The Minimum Resolution Likelihood method defines a fiducial signal region to convert ML-induced systematic effects into statistical uncertainties for unbiased signal strength estimation in collider analyses.
Parnassus releases a unified PyTorch framework offering flow-matching neural and Delphes-style parametric models for CMS, ATLAS, and ALEPH detector simulation that run on GPUs without ROOT.
Damped oscillations alleviate LN violation suppression for pseudo-Dirac HNLs, improving collider sensitivities and allowing distinction from the double-Majorana limit.
Reanalysis of CMS ttγ data shows a 2.7σ localized excess in m_ℓℓγ at 152 GeV compatible with S → W⁺W⁻γ from a narrow scalar, yielding a ratio σ(S → W⁺W⁻γ)/σ(S → W⁺W⁻) = (2.14 ± 0.77)%.
Proof-of-principle for hadron-in-fat-jet AI tagging that yields an expected 95% CL limit of B(W±→π±γ) < 2.78×10^{-5} at 450 fb^{-1}.
DNN analysis of pp → WR → ℓNR → ℓℓjj at LHC Run 2 and HL-LHC improves exclusion limits on m_WR and m_NR for unmixed, maximal-mixing, and PMNS-like scenarios over cut-based methods and probes the |Ve1|–|Vμ1| plane.
Pointwise metrics compress marginal spectra in multimodal inverse problems, and a three-part protocol using CRPS, spectrum fidelity, and calibration reverses model rankings on synthetic and particle-physics benchmarks.
Neural networks for HEP tasks can be fooled at significant rates by subtle perturbations inside uncertainty envelopes, revealing hidden systematics not captured by conventional methods.
A proposed LHC search using low-multiplicity jets plus a photon can extend sensitivity to GeV-scale particles that couple to light quarks.
Global EFT analysis of charged lepton-flavor violation at future colliders incorporates RG running, polarization, and Bayesian discrimination to identify operators, finding 10-30% shifts from tree-level mappings at multi-TeV scales.
No significant excess is observed in leptonic W/Z plus high-multiplicity soft-particle events, setting limits on Higgs to SUEP decays across a range of model parameters.
A new model with SU(2)_D symmetry and vector-like muons mediates vector dark matter, simultaneously addressing relic abundance and muon g-2 while identifying an off-resonance suppression mechanism for light DM and deriving collider bounds.
Novel multilepton signatures from Higgs decays to a light pseudoscalar decaying to e-mu pairs in type-III 2HDM can set stronger limits on LFV couplings than low-energy experiments.
Higgsformer achieves AUC 0.855 on t tbar H vs t tbar classification from raw hits, matching a Delphes-based Particle Transformer at ~40% b-tagging efficiency.
Introduces pseudo-marginal MCMC with unbiased Poisson likelihood estimator for exact inference despite noisy collider Monte Carlo simulations.
citing papers explorer
-
Proton Structure from Neural Simulation-Based Inference at the LHC
Neural simulation-based inference on unbinned top-quark pair data at 13 TeV yields improved gluon PDF precision over traditional binned analyses while incorporating experimental and theoretical uncertainties.
-
Reweighting Adversarial Networks for Unbinned Unfolding
RANs generalize moment unfolding to full phase-space unbinned unfolding via detector-level Wasserstein critics without requiring support overlap or multiple iterations.
-
Particle-Lund Multimodality in Jet Taggers
PLuM multimodal transformer improves top and H->bb jet tagging by jointly processing particle constituents and Lund plane splittings, yielding 25% higher background rejection at 25% di-Higgs efficiency.
-
Higgs Physics with the XFEL Compton $\boldsymbol{\gamma\gamma}$ Collider Concept at $\boldsymbol{\sqrt{s}=125}$ GeV
An XFEL Compton gamma-gamma collider at 125 GeV with a set transformer deep learning classifier on particle-flow point clouds can achieve high-precision Higgs measurements across hadronic, semi-leptonic, and leptonic final states including H to strange quarks.
-
Collider-Bench: Benchmarking AI Agents with Particle Physics Analysis Reproduction
Collider-Bench is a new benchmark showing that current LLM agents cannot reliably reproduce LHC analyses at the level of a physicist-in-the-loop.
-
From Information Geometry to Jet Substructure: A Triality of Cumulant Tensors, Energy Correlators, and Hypergraphs
Higher-order Fisher tensors in exponential-family coordinates of binned energy correlators are simultaneously local KL coefficients, connected cumulants, and hyperedge weights, enabling hypergraph constructions for jet substructure analysis.
-
Many Wrongs Make a Right: Leveraging Biased Simulations Towards Unbiased Parameter Inference
Template-Adapted Mixture Model uses many biased simulations for data-driven estimates of signal and background distributions, yielding unbiased signal fraction estimates with well-calibrated uncertainties.
-
Stay Positive: Neural Refinement of Sample Weights
Neural refinement of Monte Carlo sample weights via phase-space scaling and a new resampling protocol that maintains averages and uncertainties.
-
The automated computation of tree-level and next-to-leading order differential cross sections, and their matching to parton shower simulations
MadGraph5_aMC@NLO automates tree-level, NLO, shower-matched, and merged cross-section computations for collider processes in a unified flexible framework.
-
Defining a Minimum Resolution for Unbinned Analyses
The Minimum Resolution Likelihood method defines a fiducial signal region to convert ML-induced systematic effects into statistical uncertainties for unbiased signal strength estimation in collider analyses.
-
Parnassus: A GPU-enabled, Python-based Package for Fast Particle Detector Simulation and Reconstruction
Parnassus releases a unified PyTorch framework offering flow-matching neural and Delphes-style parametric models for CMS, ATLAS, and ALEPH detector simulation that run on GPUs without ROOT.
-
Lepton number violation at hadron colliders via pseudo-Dirac heavy neutral leptons
Damped oscillations alleviate LN violation suppression for pseudo-Dirac HNLs, improving collider sensitivities and allowing distinction from the double-Majorana limit.
-
Radiative Signature of New Scalar Boson Decays in the $m_{\ell \ell \gamma}$ Spectrum at the LHC
Reanalysis of CMS ttγ data shows a 2.7σ localized excess in m_ℓℓγ at 152 GeV compatible with S → W⁺W⁻γ from a narrow scalar, yielding a ratio σ(S → W⁺W⁻γ)/σ(S → W⁺W⁻) = (2.14 ± 0.77)%.
-
"Hadron-in-fat-jet'' AI Tagging to Detect Rare Decays Such as $W^{\pm}\to\pi^{\pm}\gamma$
Proof-of-principle for hadron-in-fat-jet AI tagging that yields an expected 95% CL limit of B(W±→π±γ) < 2.78×10^{-5} at 450 fb^{-1}.
-
Probing lepton flavor mixing in $W_R$ searches with machine learning at the LHC
DNN analysis of pp → WR → ℓNR → ℓℓjj at LHC Run 2 and HL-LHC improves exclusion limits on m_WR and m_NR for unmixed, maximal-mixing, and PMNS-like scenarios over cut-based methods and probes the |Ve1|–|Vμ1| plane.
-
Pointwise Metrics Mislead: An Evaluation Protocol for Multimodal Inverse Problems
Pointwise metrics compress marginal spectra in multimodal inverse problems, and a three-part protocol using CRPS, spectrum fidelity, and calibration reverses model rankings on synthetic and particle-physics benchmarks.
-
Uncovering Hidden Systematics in Neural Network Models for High Energy Physics
Neural networks for HEP tasks can be fooled at significant rates by subtle perturbations inside uncertainty envelopes, revealing hidden systematics not captured by conventional methods.
-
Low-Multiplicity Jets as Probes of GeV-Scale Light-Quark-Coupled Particles
A proposed LHC search using low-multiplicity jets plus a photon can extend sensitivity to GeV-scale particles that couple to light quarks.
-
Operator Identification in Charged Lepton-Flavor Violation: Global EFT Analysis with RG Evolution, Polarization Observables, and Bayesian Model Discrimination at Future Colliders
Global EFT analysis of charged lepton-flavor violation at future colliders incorporates RG running, polarization, and Bayesian discrimination to identify operators, finding 10-30% shifts from tree-level mappings at multi-TeV scales.
-
Search for soft unclustered energy patterns produced in association with a W or Z boson in proton-proton collisions at $\sqrt{s}$ = 13 TeV
No significant excess is observed in leptonic W/Z plus high-multiplicity soft-particle events, setting limits on Higgs to SUEP decays across a range of model parameters.
-
The Muonic Portal to Vector Dark Matter:connecting precision muon physics, cosmology, and colliders
A new model with SU(2)_D symmetry and vector-like muons mediates vector dark matter, simultaneously addressing relic abundance and muon g-2 while identifying an off-resonance suppression mechanism for light DM and deriving collider bounds.
-
Novel probes for electron-muon flavor violation from exotic Higgs decays
Novel multilepton signatures from Higgs decays to a light pseudoscalar decaying to e-mu pairs in type-III 2HDM can set stronger limits on LFV couplings than low-energy experiments.
-
Hits to Higgs: Hit-Level Higgs Classification from Raw LHC Detector Data Using Higgsformer
Higgsformer achieves AUC 0.855 on t tbar H vs t tbar classification from raw hits, matching a Delphes-based Particle Transformer at ~40% b-tagging efficiency.
-
Bring the noise: exact inference from noisy simulations in collider physics
Introduces pseudo-marginal MCMC with unbiased Poisson likelihood estimator for exact inference despite noisy collider Monte Carlo simulations.
-
Pretrained Event Classification Model for High Energy Physics Analysis
A GNN pretrained on 120M simulated HEP events generalizes to unseen processes and ATLAS data; fine-tuning boosts accuracy especially with small datasets, with CKA showing preserved encoders but altered intermediate layers.
-
Soft-Dimuon Signature from Two-Component Scalar Dark Matter at the LHC
A detector-level Monte Carlo analysis in the I(2+1)HDM reports S/B ≈ 9.8% and statistical significance S/√B = 4.93 at 4 ab⁻¹ for a benchmark two-component scalar DM point with soft dimuon invariant mass below m_Z.
-
Toponium effects on quantum steering and Bell nonlocality of top quarks
Toponium strengthens the spin-singlet component in top-quark pairs, substantially enhancing entanglement and enabling observable quantum steering at 10 sigma and Bell nonlocality at 9 sigma near threshold with current LHC data.
-
$t \bar{t}$ production as a window to invisible new physics
Phenomenological LHC study finds sensitivity to light spin-1 DM mediators in ttbar events and discrimination power from CP-sensitive angular observables in dileptonic final states.
-
Reinterpretation of ATLAS and CMS searches in monojet and mono-$V$ final states: prospects of limits on excited neutrinos
Reinterpretation of monojet and mono-V searches sets upper limits on excited-neutrino production, excluding masses up to ~4 TeV in benchmark scenarios.
-
Search for heavy Majorana neutrinos at muon-proton colliders via lepton-number-violating signals
Simulates LNV signals from heavy Majorana neutrinos (200-3000 GeV) at a 5.3 TeV μp collider and projects 2σ limits on |V_ℓN|² superior to LHC bounds for 100 fb⁻¹ and 1 ab⁻¹ luminosities.
-
Reinterpreting the ATLAS HHH$\to 6b$ Search with CheckMATE and Rivet: Validation, TRSM Benchmarks, and HL-LHC Prospects
Reimplementation and validation of ATLAS HHH→6b search in CheckMATE/Rivet with new TRSM benchmark projections and HL-LHC sensitivity estimates.
-
Probing the Rare Four-Bottom Higgs Decay $H\to b\bar b b\bar b$ at the HL-LHC and ILC
The rare Higgs decay H to four bottom quarks has a branching ratio of order 1.6e-3 with relevant destructive interference and is observable at 3-5 sigma at HL-LHC and ILC via boosted decision tree analyses in WH and ZH production.
-
Probing SMEFT Operators through $t\bar{t}t\bar{t}$ Production with Hyper-Graph Neural Networks at the LHC
A hyper-graph neural network improves discrimination of four-top production at 13 TeV, raising expected significance from 5.13 to 9.11 and enabling projected 95% CL limits on five dimension-six SMEFT Wilson coefficients at current and HL-LHC luminosities.
-
Time-dependent signals of new physics at the LHC
Incorporating timing information from time-dependent new physics signals can improve LHC search sensitivity by up to a factor of two compared to standard time-invariant analyses.
-
Kitchen Sink Anomaly Detection
A combined kitchen sink observable set of Energy Flow Polynomials and subjettiness variables outperforms standard baselines in sensitivity to a wide range of resonant signals, with new public benchmarks released and an attribute bagging variant reducing training cost.
-
Probing Freeze-In Dark Matter via a Spin-2 Portal at the LHC with Vector Boson Fusion and Machine Learning
LHC vector boson fusion searches enhanced by machine learning can probe substantial regions of cosmologically viable parameter space for freeze-in dark matter mediated by a spin-2 particle.
-
Validating a Machine Learning Approach to Identify Quenched Jets in Heavy-Ion Collisions
An LSTM model trained on simulated jet substructure learns to predict true jet energy loss and distinguishes quenching signatures even after realistic detector effects are applied.
-
Stopping Dark Mesons in Their Tracks with Long-Lived Particle and Resonant Signatures
Recast LHC searches yield a ~1.2 TeV lower bound on long-lived charged dark mesons and show that anomaly-driven diboson resonances can reconstruct UV parameters like dark flavor and color numbers from IR measurements.
-
Optimizing The Cut And Count Method In Phenomenological Studies
An iterative ranking-based optimization of cut-and-count using MadAnalysis5 enhances signal-background separation and discovery reach for singly charged Higgs in the Two Higgs Doublet Model.
-
Higgs Assisted Razor Search for Higgsinos at a 100 TeV pp Collider
A Higgs-assisted razor search with ML for Higgsino NLSPs decaying to Bino LSP via Z/h at 100 TeV projects 5σ discovery to 1.4 TeV Higgsino mass (Bino ~0.9 TeV) with 3000 fb^{-1}.
-
New Avenues of Heavy Neutral Lepton at Muon Collider
In a U(1) gauged seesaw model, Z'Z' fusion processes at muon colliders enable HNL pair production via Z'Z'→H→NN and Z'Z'→NN, yielding LNV signals not suppressed by Higgs mixing.
-
Freeze-in at all couplings
In low-reheating-temperature charged-parent freeze-in dark matter models, stronger couplings are viable if both dark matter and mediator number densities are tracked, with updated LHC and lepton-flavor-violation constraints.
-
Probing TeV-Scale Inverse-Seesaw Leptogenesis and Majorana Dark Matter in $U(1)_{B-L}$ Models at Multi-TeV Muon Colliders
Inverse-seesaw U(1)_{B-L} model correlates leptogenesis, Majorana DM relic density, and neutrino masses with collider signatures in dilepton and single-lepton channels.
-
Probing anomalous quartic gauge couplings in same-sign $W$ boson scattering with polarization and spin correlation
Spin-correlation asymmetries in same-sign WW production yield sensitivity to anomalous WWWW couplings comparable to transverse-mass distributions, and their combination improves Wilson-coefficient limits while respecting unitarity cuts.
-
Probing Singlet Vector-Like Top Quarks in the Hadronic tZ Channel at the HL-LHC using Machine and Deep Learning Architectures
Monte Carlo simulation plus XGBoost and GNN classification yields 2σ exclusion and 5σ discovery contours in the (g*, m_T) plane for a singlet vector-like top quark at 3000 fb⁻¹ in the hadronic tZ channel.
-
Deep Neural Networks for Heavy Lepton-Flavor-Violating Higgs Searches at the LHC
DNN classifiers with mass-dependent thresholds reduce expected 95% CL upper limits on H to mu tau cross sections by 36-46% versus collinear mass baseline, while a regression network improves mass resolution by up to 21%.
-
Phenomenology of electroweak spin-1 resonances
Composite Higgs models with SU(2)_L × SU(2)_R predict spin-1 resonances mixing with electroweak bosons that remain viable at the LHC down to masses of about 1.5 TeV.
-
Same-Sign Tetralepton Signature at $\mu$TRISTAN
The paper identifies promising parameter regions for observing same-sign tetralepton events from charged Higgs pair and single production decaying to muons and heavy neutral leptons at μTRISTAN.
-
Probing Flavor-Violating Higgs Decays in the Type-III Two-Higgs-Doublet Model at the LHC and HL-LHC
In the Type-III 2HDM, neutral and heavy charged flavor-violating Higgs decays can exceed 5 sigma significance at 300 fb^{-1} luminosity while the light charged mode is more background-limited.
-
Sensitivity to top-quark FCNC interactions at future muon colliders
A 10 TeV muon collider with 10 ab^{-1} could reach O(10^{-3}) sensitivity on top-quark FCNC couplings, improving current ATLAS/CMS bounds by more than an order of magnitude.