TY  - JOUR
PB  - SciPost Foundation
DO  - 10.21468/SciPostPhys.19.4.108
TI  - A Lorentz-equivariant transformer for all of the LHC
PY  - 2025/10/23
UR  - https://www.scipost.org/SciPostPhys.19.4.108
JF  - SciPost Physics
JA  - SciPost Phys.
VL  - 19
IS  - 4
SP  - 108
A1  - Brehmer, Johann
AU  - Bresó, Victor
AU  - de Haan, Pim
AU  - Plehn, Tilman
AU  - Qu, Huilin
AU  - Spinner, Jonas
AU  - Thaler, Jesse
AB  - We show that the Lorentz-Equivariant Geometric Algebra Transformer (L-GATr) yields state-of-the-art performance for a wide range of machine learning tasks at the Large Hadron Collider. L-GATr represents data in a geometric algebra over space-time and is equivariant under Lorentz transformations. The underlying architecture is a versatile and scalable transformer, which is able to break symmetries if needed. We demonstrate the power of L-GATr for amplitude regression and jet classification, and then benchmark it as the first Lorentz-equivariant generative network. For all three LHC tasks, we find significant improvements over previous architectures.
ER  -

@Article{10.21468/SciPostPhys.19.4.108,
	title={{A Lorentz-equivariant transformer for all of the LHC}},
	author={Johann Brehmer and Víctor Bresó and Pim de Haan and Tilman Plehn and Huilin Qu and Jonas Spinner and Jesse Thaler},
	journal={SciPost Phys.},
	volume={19},
	pages={108},
	year={2025},
	publisher={SciPost},
	doi={10.21468/SciPostPhys.19.4.108},
	url={https://scipost.org/10.21468/SciPostPhys.19.4.108},
}