Data Science Techniques for the Standard Model Effective Field Theory in the search of New Physics at High Energy Colliders

The focus of this project is on New Physics searches at the Large Hadron Collider (LHC) and future colliders. We aim to combine two growing fields within High Energy Physics: Effective Field Theories (EFTs) and Data Science. Our goal is to study a series of electroweak processes involving the Higgs boson, which are well known within the Standard Model (SM) at LHC but have not been calculated previously within the Standard Model Effective Field Theory (SMEFT). We will investigate different multivariate analysis (MVA) frameworks to study these EFT parametrisations at LHC since they depend on a large number of uncorrelated variables. We will focus on Matrix Element (ME) techniques, but not discarding a priori other MVA techniques such as boosted-decision-trees (BDTs) or Artificial Neural Networks (ANNs). An integral part involves developing a public code to host such MVA-EFT algorithm, a code that is useful for the different LHC collaborations to use in their experimental analysis. To perform the last two steps we will work closely with the experimental groups at LHC and we will benefit from solid collaborations with international experts in the field of data science (from Durham, Nikhef and Valencia) as well as with the local experimental groups. Lastly, we will adapt our tool to the next generation of large scale computing hardware. These objectives altogether represent an innovative strategy to tackle New Physics searches at LHC from a model independent yet experiment-based point of view. This approach includes the novel feature that its tools will be designed to work at the experimental analysis level, in contrast to the common approach by which the SMEFT studies are performed on published results.