Erlangen 2026 – scientific programme

Parts | Days | Selection | Search | Updates | Downloads | Help

T: Fachverband Teilchenphysik

T 71: Data, AI, Computing, Electronics VII

T 71.4: Talk

Thursday, March 19, 2026, 17:00–17:15, KH 00.024

A Machine-Learning based Topological Algorithm for the Level-1 Trigger System of CMS — •Lukas Ebeling, Johannes Haller, Artur Lobanov, and Matthias Schröder — Institut für Experimentalphysik, Universität Hamburg

At CMS, the level-1 trigger (L1T) system is crucial to select events of interest in order to keep the data-taking rate at a level that can be processed by the readout and storage system. We present a machine-learning (ML) based algorithm for the L1T system, designed to identify di-Higgs (HH) production events. The algorithm leverages the full event topology and improves the HH signal efficiency at low p_T compared to previous single-object based triggers. Despite being constrained in architecture by strict latency requirements and limited FPGA hardware, the ML trigger achieves high signal efficiencies while maintaining acceptable rates. The talk will highlight the achieved trigger performance, discuss the integration into the CMS software framework as well as the development of a realistic trigger scenario for running in 2026 collisions.

Keywords: Topological Trigger; Level-1 Trigger; CMS; di-Higgs