Simulation-based inference in the search for CP violation in leptonic WH production

Sources of CP violation beyond the Standard Model (BSM) are required to explain the baryonic asymmetry of the Universe. In this work, we studied BSM CP violating components in the HWW interaction in WH production, parametrized by an effective dimension-6 CP-odd operator. We studied the feasibility of using a machine learning simulation-based inference method, SALLY, that estimates a detector-level optimal observable, and compared its sensitivity to that of energy-dependent and angular observables, also studying various binnings for these observables. We show that, when linear effects in the Wilson coefficients dominate, an observable such as SALLY will lead to optimal limits. In the regions of phase space where quadratic effects are significant, the best sensitivity comes from a 2D combination of an angular observable and an energy-dependent variable with bins in the high-energy region. This works aims at testing the current multivariate techniques in view of informing analysis strategies for LHC Run 3.