Jet Single Shot Detection
| Autor: | Adrian Alan Pol, Anat Klempner, Olya Sirkin, Jennifer Ngadiuba, Maurizio Pierini, Roi Halily, Vladimir Loncar, Katya Govorkova, Thea Klaeboe Aarrestad, Sioni Summers, Tal Kopetz |
|---|---|
| Rok vydání: | 2021 |
| Předmět: |
Jet (fluid)
Large Hadron Collider Calorimeter (particle physics) hep-ex 010308 nuclear & particles physics Computer science Physics QC1-999 FOS: Physical sciences 01 natural sciences Convolutional neural network Measure (mathematics) Object detection High Energy Physics - Experiment High Energy Physics - Experiment (hep-ex) 0103 physical sciences High Energy Physics::Experiment 010306 general physics Focus (optics) Scaling Algorithm Particle Physics - Experiment |
| Zdroj: | EPJ Web of Conferences, Vol 251, p 04027 (2021) |
| ISSN: | 2100-014X |
| DOI: | 10.1051/epjconf/202125104027 |
| Popis: | We apply object detection techniques based on Convolutional Neural Networks to jet reconstruction and identification at the CERN Large Hadron Collider. In particular, we focus on CaloJet reconstruction, representing each event as an image composed of calorimeter cells and using a Single Shot Detection network, called Jet-SSD. The model performs simultaneous localization and classification and additional regression tasks to measure jet features. We investigate TernaryWeight Networks with weights constrained to {-1, 0, 1} times a layer- and channel-dependent scaling factors. We show that the quantized version of the network closely matches the performance of its full-precision equivalent. We apply object detection techniques based on Convolutional Neural Networks to jet reconstruction and identification at the CERN Large Hadron Collider. In particular, we focus on CaloJet reconstruction, representing each event as an image composed of calorimeter cells and using a Single Shot Detection network, called Jet-SSD. The model performs simultaneous localization and classification and additional regression tasks to measure jet features. We investigate Ternary Weight Networks with weights constrained to {-1, 0, 1} times a layer- and channel-dependent scaling factors. We show that the quantized version of the network closely matches the performance of its full-precision equivalent. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|