Semi-visible jets, energy-based models, and self-supervision
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 a...
Saved in:
| Main Authors: | , , , , |
|---|---|
| Format: | Article (Journal) |
| Language: | English |
| Published: |
3 February 2025
|
| In: |
SciPost physics
Year: 2025, Volume: 18, Issue: 2, Pages: 1-17 |
| ISSN: | 2542-4653 |
| DOI: | 10.21468/SciPostPhys.18.2.042 |
| Online Access: | Verlag, lizenzpflichtig, Volltext: https://doi.org/10.21468/SciPostPhys.18.2.042 Verlag, lizenzpflichtig, Volltext: https://scipost.org/10.21468/SciPostPhys.18.2.042 
|
| Author Notes: | Luigi Favaro, Michael Krämer, Tanmoy Modak, Tilman Plehn and Jan Rüschkamp |
| Summary: | 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. |
|---|---|
| Item Description: | Gesehen am 25.07.2025 |
| Physical Description: | Online Resource |
| ISSN: | 2542-4653 |
| DOI: | 10.21468/SciPostPhys.18.2.042 |