Speaker
Raghav Kansal
(Caltech & Fermilab)
Description
We describe several novel machine learning techniques to improve CMS searches and measurements. These include state-of-the-art transformer models for hadronic jet classification, flow and decorrelation methods for background estimation and MC parameter reweighting, and end-to-end analysis optimization. The impact of these advances on a selection of recent CMS results will be discussed.
Primary author
Raghav Kansal
(Caltech & Fermilab)
