A Measurement of Boosted Dibosons with Gaussian Process Background Modeling at the ATLAS Detector

This dissertation presents a measurement of $WW$, $WZ$, and $ZZ$ production in the semileptonic final state at the ATLAS detector. The measurement was performed using $139;\text{fb}^{-1}$ of $\sqrt{s} = 13 \tev$ proton-proton collisions collected at the Large Hadron Collider between 2015 and 2018. An innovative data-driven approach to model the V+jets background was investigated, using a Bayesian method called Gaussian process regression. The method, more commonly found in other disciplines and yet to see much adoption in high-energy physics, was established to be much more robust and precise than traditional approaches using functional fits. As the analysis is still blinded as of this publication, expected results are presented here instead. Differential fiducial cross sections are obtained in several variables, and a total sensitivity of around $10 \sigma$ to the Standard Model process is expected. Expected limits on the Effective Field Theory coefficient $c_W$ are determined at approximately $|c_W/\Lambda^2| < 0.42 \text{ TeV}^{-2}$ at the linear order and $|c_W/\Lambda^2| < 0.03 \text{ TeV}^{-2}$ at the quadratic order. In addition to the diboson measurement, firmware, software, and algorithm developments on tracking in the trigger for the HL-LHC upgrade are presented. The highly increased pileup of the HL-LHC environment requires a much smarter trigger system to reject background events, and tracking enables more informed decisions. Two algorithms, a pattern matching method and a Hough transform approach, were implemented in a simulation framework and optimized for the HL-LHC environment.