Tanasini, M. (2023) Using Graph Neural Networks to improve jet flavour-tagging and its modeling for the ATLAS experiment at the LHC. Il nuovo cimento C, 46 (3). pp. 1-10. ISSN 1826-9885
|
Text
ncc12626.pdf - Published Version Download (1MB) | Preview |
Abstract
Graph Neural Networks (GNNs) are machine-learning algorithms particularly suitable for modeling data with complex topological correlations. In this communication, their application to jet flavour-tagging for the ATLAS experiment at the Large Hadron Collider is presented. A new GNN algorithm to identify jets containing heavy-flavour hadrons by representing them as graphs of tracks and silicon hits is illustrated. The performance of the modern flavour-tagging algorithms poses challenges when applied on simulated events containing multiple jets, as they reduce the statistical precision of the simulated samples. To overcome this, a GNNbased technique that increases the statistical power of the samples by weighting events based on their likelihood of containing flavour-tagged jets is described.
| Item Type: | Article |
|---|---|
| Subjects: | 500 Scienze naturali e Matematica > 530 Fisica |
| Depositing User: | Marina Spanti |
| Date Deposited: | 29 Jan 2024 12:03 |
| Last Modified: | 29 Jan 2024 12:03 |
| URI: | http://eprints.bice.rm.cnr.it/id/eprint/22593 |
Actions (login required)
 |
View Item |
