TY  - JOUR
PB  - SciPost Foundation
DO  - 10.21468/SciPostPhys.18.2.042
TI  - Semi-visible jets, energy-based models, and self-supervision
PY  - 2025/02/03
UR  - https://scipost.org/SciPostPhys.18.2.042
JF  - SciPost Physics
JA  - SciPost Phys.
VL  - 18
IS  - 2
SP  - 042
A1  - Favaro, Luigi
AU  - Krämer, Michael
AU  - Modak, Tanmoy
AU  - Plehn, Tilman
AU  - Rüschkamp, Jan
AB  - We present DarkCLR, a novel framework for detecting semi-visible jets at the LHC. DarkCLR uses a self-supervised contrastive-learning approach to create observables that are approximately invariant under relevant transformations. We use background-enhanced data to create a sensitive representation and evaluate the representations using a CLR-inspired anomaly score and a normalized autoencoder as density estimators. Our results show a remarkable sensitivity for a wide range of semi-visible jets and are more robust than a supervised classifier trained on a specific signal.
ER  -

@Article{10.21468/SciPostPhys.18.2.042,
	title={{Semi-visible jets, energy-based models, and self-supervision}},
	author={Luigi Favaro and Michael Krämer and Tanmoy Modak and Tilman Plehn and Jan Rüschkamp},
	journal={SciPost Phys.},
	volume={18},
	pages={042},
	year={2025},
	publisher={SciPost},
	doi={10.21468/SciPostPhys.18.2.042},
	url={https://scipost.org/10.21468/SciPostPhys.18.2.042},
}