arXiv preprint ·
QGrAT: Quantum Grade-Aware Transformer for Jet Tagging
- Published
- Source
- arXiv preprint, posted before peer review
- Original
- Read it at arXiv
- Topics
Abstract
Lorentz symmetry and the three-prong structure of hadronic top decays provide physical guidance for designing compact quantum jet taggers. We introduce QGrAT (Quantum Grade-Aware Transformer), a quantum transformer that incorporates these structures into its interactions between constituent qubits. Geometric features constructed from the constituent four-momenta are organised by the grades of the spacetime Clifford algebra $C\ell(1,3)$, and their grade-specific inner products determine the pairwise attention couplings. For fixed constituents and co-transforming reference vectors, these couplings are Lorentz invariant, although the component-wise input encoding makes the complete classifier frame dependent. An auxiliary qubit introduces an explicit three-particle interaction through products of pairwise Minkowski inner products, while a trainable in-circuit readout converts the resulting quantum state into a classification score. We evaluate the architecture in state-vector simulation on the Top Tagging benchmarks. With ten constituents per jet, the configuration using grades $0\text{-}3$ and the three-particle ancilla requires twelve qubits and $144$ trainable parameters, reaching AUC $0.944 \pm 0.002$ and background rejection $1/\varepsilon_B = 35.7 \pm 2.4\ (96.8 \pm 7.7)$ at signal efficiency $\varepsilon_S = 0.50\ (0.30)$. This performance is competitive with the L-GATr-Slim baseline at matched constituent and parameter budgets. Adding the three-particle ancilla raises background rejection at $50\%$ signal efficiency from $19.9$ to $35.7$, whereas extending the representation to the tested grade-$4$ feature gives no consistent improvement. The study establishes a framework for incorporating spacetime geometry into quantum attention and identifies explicit three-particle correlations as the principal source of improved background rejection among the architectural variations tested.
The authors’ abstract, as arXiv lists it. arXiv’s listings, abstracts included, are free for anyone to reuse.
Read it on ReadSearch
Each opens after you create an account, and brings you back to this paper.
More in High Energy Physics - Phenomenology
Vortex structures in momentum distributions from multi-pair production
Stabilizer Rényi Entropy Without a Phase Reference
Indirect Dark Matter Detection in the Multi-Messenger Era: Experiments, Systematics, Anomalies, and Prospects for Discovery
Quantum-enhanced Search for Galactic Dark Photons Using Cat-State Cavities
Can bosenovae deplete the galactic dark matter?
Get the few papers worth your time in High Energy Physics - Phenomenology every morning.
ReadSearch reads the new papers in your field each day and brings you the few worth reading, each with its key finding in a sentence.