中国物理学会高能物理分会第十三届全国粒子物理学术会议（2021）

The identification of massive particles decaying into bottom quark pairs is important for the physics program of the ATLAS experiment at the Large Hadron Collider. A neural network (NN) based double b-tagging algorithm named X→bb ̅ tagger is developed and calibrations of the tagger are performed using proton-proton collision data corresponding to 139〖fb〗^(-1) collected at a centre-of-mass energy of √s=13TeV. The technique of X→bb ̅ mis-tag rate calibration is developed based on semi-leptonic decay tt ̅ events which have rich non- bb ̅ flavor combination and high statistics. The mis-tag efficiency is measured and the scale factor is found to be in a range of 1~1.1 with uncertainty less than 16%.