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

arXiv:2605.18382v1 Announce Type: cross Abstract: We present a phenomenological study of tar{t}tar{t} production in proton-proton collisions at sqrt{s} = 13TeV, using a Hyper-Graph Neural Network (H-GNN) to discriminate multilepton signal events from the dominant SM backgrounds, namely tar{t}W, tar{t}Z, tar{t}H, tar{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 tar{t}tar{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 tar{t}tar{t} signal and yields a statistical significance of Z = 9.11 at an integrated luminosity of L = 140fb^{-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 O_{Phi u}, O^{(1)}{tt}, O^{(1)}{qq}, O^{(1)}{qt}, and O^{(8)}{qt}, and we project the expected sensitivity at the HL-LHC integrated luminosities of 1000fb^{-1} and 3000fb^{-1} with 50% uncertainty on the background estimation.

Source: arXiv cs.AI | 2026-05-19