TY  - JOUR
PB  - SciPost Foundation
DO  - 10.21468/SciPostPhys.20.2.038
TI  - BitHEP — The limits of low-precision ML in HEP
PY  - 2026/02/10
UR  - https://scipost.org/SciPostPhys.20.2.038
JF  - SciPost Physics
JA  - SciPost Phys.
VL  - 20
IS  - 2
SP  - 038
A1  - Krause, Claudius
AU  - Wang, Daohan
AU  - Winterhalder, Ramon
AB  - The increasing complexity of modern neural network architectures demands fast and memory-efficient implementations to mitigate computational bottlenecks. In this work, we evaluate the recently proposed Bitnet architecture in HEP applications, assessing its performance in classification, regression, and generative modeling tasks. Specifically, we investigate its suitability for quark-gluon discrimination, SMEFT parameter estimation, and detector simulation, comparing its efficiency and accuracy to state-of-the-art methods. Our results show that while Bitnet consistently performs competitively in classification tasks, its performance in regression and generation varies with the size and type of the network, highlighting key limitations and potential areas for improvement.
ER  -

@Article{10.21468/SciPostPhys.20.2.038,
	title={{BitHEP — The limits of low-precision ML in HEP}},
	author={Claudius Krause and Daohan Wang and Ramon Winterhalder},
	journal={SciPost Phys.},
	volume={20},
	pages={038},
	year={2026},
	publisher={SciPost},
	doi={10.21468/SciPostPhys.20.2.038},
	url={https://scipost.org/10.21468/SciPostPhys.20.2.038},
}