Learning How to Count: A High Multiplicity Search for the LHC

44 citations (42 excluding self-citations).

The Problem

High-multiplicity BSM signals, such as those from cascade decays or RPV SUSY, produce so many jets that standard counting strategies break down. Our earlier work on fat jet masses showed that reclustering into large-radius jets and cutting on jet mass was a significant improvement, but the jet mass variable treats a fat jet containing three hard subjets the same as one containing six. More granular information about the internal structure of fat jets could provide additional discrimination against QCD backgrounds.

The Key Idea

Instead of just measuring the mass of fat jets, count the number of subjets inside each one. The total subjet count across all fat jets in an event serves as a proxy for the true parton multiplicity, but is far more robust than counting resolved narrow jets because the large-radius clustering is insensitive to soft radiation and pileup. We developed two subjet-counting algorithms and combined the subjet count with summed jet mass and missing energy cuts. Across eight diverse benchmark signals (cascade decays, three-body decays, multi-top final states), subjet counting improved sensitivity compared to the fat jet mass approach while reducing reliance on missing energy, and the method naturally supports data-driven QCD background estimation.

Recollections

[To be added.]

• DOI
• arXiv:1302.1870
• INSPIRE