Transformer-based Flavour Tagging in ATLAS

16 Jul 2026, 09:50
12m
深圳厅 (2号楼二楼)

深圳厅

2号楼二楼

Speaker

Bingxuan Liu (SunYat-sen University)

Description

The ATLAS collaboration has released its next-generation flavour tagging algorithm, based on a transformer architecture [1]. The performance is enhanced significantly and the performance in data matches that in the simulation. Recent physics analyses have reported large improvements in sensitivity thanks to the new tagger [2]. In this talk, I will discuss how the tagger was developed and the stories behind.

[1] Nature Commun. 17 (2026) 541
[2] Phys. Lett. B 876 (2026) 140280

请选择分会 TeV物理和超出标准模型新物理

Primary author

Bingxuan Liu (SunYat-sen University)

Presentation materials

There are no materials yet.