Speaker
Jia Fu Low
(University of Florida)
Description
"In order to preserve its ability to do physics at the electroweak scale in the HL-LHC era, the CMS experiment has to maintain low trigger thresholds that are robust against the large number of interactions per bunch crossing expected at the HL-LHC. Specifically,
the Level-1 (L1) reconstruction algorithms for prompt muon triggers need significant improvement. Moreover, there is an emerging strong interest in new physics signatures that involve long-lived particles, resulting in the production of highly displaced muons. Machine Learning techniques have been shown to provide significant gains in the performance of offline reconstruction algorithms, and with recent advancements in FPGA technology, some of these algorithms can now be implemented in the L1 hardware.
We present here novel techniques to reconstruct both prompt and displaced Level-1 endcap muons using Artificial Neural Network algorithms executed in FPGAs. The presentation will describe the approach we’ve adopted, demonstrating its performance, and show its current implementation in firmware. Plans for larger scale demonstration and integration into the CMS L1 trigger will also be presented. “
Primary authors
Jia Fu Low
(University of Florida)
Sergo Jindariani
(Fermilab)
