Fast $b$-tagging at the high-level trigger of the ATLAS experiment in LHC Run 3

The ATLAS experiment relies on real-time hadronic jet reconstruction and$b$-tagging to record fully hadronic events containing $b$-jets. Thesealgorithms require track reconstruction, which is computationally expensive andcould overwhelm the high-level-trigger farm, even at the reduced event ratethat passes the ATLAS first stage hardware-based trigger. In LHC Run 3, ATLAShas mitigated these computational demands by introducing a fastneural-network-based $b$-tagger, which acts as a low-precision filter usinginput from hadronic jets and tracks. It runs after a hardware trigger andbefore the remaining high-level-trigger reconstruction. This design relies onthe negligible cost of neural-network inference as compared to trackreconstruction, and the cost reduction from limiting tracking to specificregions of the detector. In the case of Standard Model $HH \rightarrowb\bar{b}b\bar{b}$, a key signature relying on $b$-jet triggers, the filterlowers the input rate to the remaining high-level trigger by a factor of fiveat the small cost of reducing the overall signal efficiency by roughly 2%.