Speaker
Description
The measurement of beauty-jet (b-jet) production in proton–proton (pp) collisions provides a stringent test of perturbative Quantum Chromodynamics (pQCD), owing to the large mass of the beauty quark, and serves as an essential baseline for b-jet studies in heavy-ion collisions. In particular, the excellent tracking and vertexing capabilities of the ALICE detector enable access to the low transverse-momentum ($p_\mathrm{T}$) region of b-jet production, where non-perturbative QCD effects become significant and theoretical uncertainties are large. Furthermore, the study of b-jet quenching in the low-$p_\mathrm{T}$ region of heavy-ion collisions is crucial for investigating the mass dependence of energy-loss mechanisms in the hot and dense QCD medium.
However, in heavy-ion collisions, the large combinatorial background significantly degrades the performance of conventional b-jet tagging methods based on impact-parameter or secondary-vertex reconstruction. The Graph Neural Network (GNN)-based b-jet tagging method shows a substantially superior performance compared to conventional methods in pp collisions and is expected to outperform the classical approaches also in heavy-ion collisions, despite any degradation due to background. Consequently, the implementation of a GNN-based method in ALICE is expected to enable low-$p_\mathrm{T}$ b-jet measurements in heavy-ion collisions with improved precision.
This poster presents a measurement of the b-jet cross section in pp collisions at $\sqrt{s}=13.6~\mathrm{TeV}$ using the GNN-based tagging method. The improved tagging performance of the GNN approach is discussed, together with future prospects for b-jet analyses in ALICE based on this method.
| Is this an experimental talk? | Yes |
|---|---|
| Is this on behalf of a collaboration? | Yes |
| Which collaboration? | ALICE |
| Are you willing to present as a poster if it is not selected for oral presentation? | Yes |