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.
ParticleNet: Jet Tagging via Particle Clouds,
17 Pith papers cite this work. Polarity classification is still indexing.
17
Pith papers citing it
citation-role summary
background 3
method 1
citation-polarity summary
representative citing papers
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.
LP2B encoding converts Lund plane jet representations into Bloch sphere qubit states, enabling a QTTN that matches classical LundNet performance on polarization tagging and W/top tagging with three orders of magnitude fewer parameters and improved low-data regime results.
A physics-guided neural network embedding AdS5 Dirac equation and holographic Pomeron fits SLAC proton F2 data with chi-squared per degree of freedom of 0.91 and identifies a kinematic crossover at x approximately 0.19 while recovering Pomeron intercept of 1.0786.
IAFormer uses boost-invariant pairwise quantities and differential attention to create a sparse Transformer that achieves state-of-the-art classification on top-quark and quark-gluon jet datasets while using over an order of magnitude fewer parameters than prior Particle Transformer models.
First NLO-QCD amplitude-assisted ML regression for longitudinal-boson production rate in di-boson events at the LHC, benchmarked against random forests.
Transformer models trained on FCC-ee CLD simulation reconstruct hadronic taus with per-mille mis-ID, F1 up to 0.95, sub-per-mille charge errors, and percent-level transverse momentum resolution.
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.
Geometric deep learning provides a unified mathematical framework based on grids, groups, graphs, geodesics, and gauges to explain and extend neural network architectures by incorporating physical regularities.
Tracking detectors achieve 1σ mechanism discrimination in neutrinoless double-beta decay with a few events and 3σ with roughly 10 events, rising to about 25 events under realistic reconstruction uncertainties.
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.
Explainability techniques applied to LundNet show that assigned node importance correlates with classical jet substructure observables such as N-subjettiness ratios and energy correlation functions, with shifts across transverse-momentum regimes.
An edge-weighted multi-graph GNN (E-PCN/KIGNet) that encodes four jet kinematic variables improves JetClass jet-tagging accuracy and attributes most predictions to angular separation and transverse momentum.
Deep learning on all particles via holistic analysis and Advanced Color Singlet Identification improves Higgs signal extraction up to sixfold in high-energy collisions.
Adversarial training enhances robustness of jet tagging classifiers while preserving performance, with loss surface geometry providing insights into correlations and vulnerability.
FCC-ee projections indicate at least 10 times better precision on the H to tau tau cross-section than the LHC through ZH and VBF channels plus improved tau reconstruction methods.
citing papers explorer
-
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.
-
Lund Plane to Bloch (LP2B) Encoding for Object and Polarization Tagging with Quantum Jet Substructure
LP2B encoding converts Lund plane jet representations into Bloch sphere qubit states, enabling a QTTN that matches classical LundNet performance on polarization tagging and W/top tagging with three orders of magnitude fewer parameters and improved low-data regime results.
-
Probing Proton Structure via Physics-Guided Neural Networks in Holographic QCD
A physics-guided neural network embedding AdS5 Dirac equation and holographic Pomeron fits SLAC proton F2 data with chi-squared per degree of freedom of 0.91 and identifies a kinematic crossover at x approximately 0.19 while recovering Pomeron intercept of 1.0786.
-
IAFormer: Interaction-Aware Transformer network for collider data analysis
IAFormer uses boost-invariant pairwise quantities and differential attention to create a sparse Transformer that achieves state-of-the-art classification on top-quark and quark-gluon jet datasets while using over an order of magnitude fewer parameters than prior Particle Transformer models.
-
Higher-order effects in amplitude-assisted polarisation extraction with machine-learning techniques
First NLO-QCD amplitude-assisted ML regression for longitudinal-boson production rate in di-boson events at the LHC, benchmarked against random forests.
-
ParticleTransformer is all you need for reconstructing hadronic tau leptons
Transformer models trained on FCC-ee CLD simulation reconstruct hadronic taus with per-mille mis-ID, F1 up to 0.95, sub-per-mille charge errors, and percent-level transverse momentum resolution.
-
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.
-
Geometric Deep Learning: Grids, Groups, Graphs, Geodesics, and Gauges
Geometric deep learning provides a unified mathematical framework based on grids, groups, graphs, geodesics, and gauges to explain and extend neural network architectures by incorporating physical regularities.
-
Statistical sensitivity of neutrinoless double-beta decay exchange mechanism discrimination by tracking experiments
Tracking detectors achieve 1σ mechanism discrimination in neutrinoless double-beta decay with a few events and 3σ with roughly 10 events, rising to about 25 events under realistic reconstruction uncertainties.
-
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.
-
Explainable AI for Jet Tagging: A Comparative Study of GNNExplainer, GNNShap, and GradCAM for Jet Tagging in the Lund Jet Plane
Explainability techniques applied to LundNet show that assigned node importance correlates with classical jet substructure observables such as N-subjettiness ratios and energy correlation functions, with shifts across transverse-momentum regimes.
-
KIGNet: Physics-Motivated Multi-Graph Representation Learning for Explainable Jet Tagging
An edge-weighted multi-graph GNN (E-PCN/KIGNet) that encodes four jet kinematic variables improves JetClass jet-tagging accuracy and attributes most predictions to angular separation and transverse momentum.
-
Learning from all particles in high-energy collisions
Deep learning on all particles via holistic analysis and Advanced Color Singlet Identification improves Higgs signal extraction up to sixfold in high-energy collisions.
-
Improving robustness of jet tagging algorithms with adversarial training: exploring the loss surface
Adversarial training enhances robustness of jet tagging classifiers while preserving performance, with loss surface geometry providing insights into correlations and vulnerability.
-
Projections of H$\to\tau\tau$ cross-section at FCC-ee
FCC-ee projections indicate at least 10 times better precision on the H to tau tau cross-section than the LHC through ZH and VBF channels plus improved tau reconstruction methods.
- Looking inside jets: an introduction to jet substructure and boosted-object phenomenology