PreprintCC BY 4.0arXiv
Lorentz-Equivariant Geometric Algebra Transformers for Top-Quark Event Reconstruction
Abstract
Results are presented for Lorentz-equivariant geometric algebra transformers (LGaTr) in three event reconstruction tasks involving top quarks. Geometric algebra provides a common Lorentz-covariant representation of four-momenta and typed physical objects. Performance of LGaTr is evaluated on three different physics tasks - tt̅/tWb separation, FCNC detection in single top-quark production and universal event reconstruction. LGaTr is outperforming the baseline networks in tW/tt̅ separation and full event reconstruction and is showing comparable or better performance to standard transformer-based networks. The implementation is provided in a companion repository https://github.com/emil2001/top-quark-universal-nets.
Related
- Particle transformers for identifying Lorentz-boosted Higgs bosons decaying to a pair of W bosons
- Projective relativistic dynamics from a quantum Lorentz algebra: time-dependent deformed Hamiltonians on (anti-)de Sitter spaces
- Restricting Trainable Lie-Algebra Growth in Equivariant Quantum Networks via Hierarchical Ancilla-Controlled Subspace Projections
- End-to-end Particle Reconstruction with Transformers in High-Granularity Calorimeters
- Projective Geometric Algebra Graph Neural Networks: Edge-Modulated Message Passing via the Geometric Product
BibTeX
@misc{abasov2026lorentzequivariant,
title = {{Lorentz-Equivariant Geometric Algebra Transformers for Top-Quark Event Reconstruction}},
author = {E. Abasov and L. Dudko and A. Markina and M. Perfilov and P. Volkov and G. Vorotnikov and A. Zaborenko},
howpublished = {arXiv},
year = {2026},
month = oct,
doi = {10.48550/arxiv.2610.29678},
eprint = {2610.29678},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2610.29678},
note = {CC BY 4.0},
}
RIS
TY - UNPB
TI - Lorentz-Equivariant Geometric Algebra Transformers for Top-Quark Event Reconstruction
AU - E. Abasov
AU - L. Dudko
AU - A. Markina
AU - M. Perfilov
AU - P. Volkov
AU - G. Vorotnikov
AU - A. Zaborenko
PY - 2026
DA - 2026/10/21
PB - arXiv
DO - 10.48550/arxiv.2610.29678
UR - https://arxiv.org/abs/2610.29678
AB - Results are presented for Lorentz-equivariant geometric algebra transformers (LGaTr) in three event reconstruction tasks involving top quarks. Geometric algebra provides a common Lorentz-covariant representation of four-momenta and typed physical objects. Performance of LGaTr is evaluated on three different physics tasks - tt̅/tWb separation, FCNC detection in single top-quark production and universal event reconstruction. LGaTr is outperforming the baseline networks in tW/tt̅ separation and full event reconstruction and is showing comparable or better performance to standard transformer-based networks. The implementation is provided in a companion repository https://github.com/emil2001/top-quark-universal-nets.
ER -