Search for Higgs Pair Production in bbγγ Final States

Measuring the Higgs boson's trilinear self-coupling λ is crucial, as it offers insights into the shape of the scalar potential and helps us understand the dynamic origins of the Higgs mechanism. At the LHC, this self-coupling can be directly observed through Higgs boson pair (HH) production. Additionally, many theories beyond the Standard Model propose the existence of heavy particles that couple to pairs of Higgs bosons, making this an ideal area for exploring new physics. We focus on both resonant and non-resonant HH productions in the bbγγ final state (pp→HH→bbγγ). The HH→bbγγ channel is particularly sensitive to self-coupling since the H → bb decay has a high branching ratio, while H → γγ provides a high-purity sample with two energetic photons. We will present preliminary analysis results from CMS Run 3 data and Monte Carlo simulations, including background estimation, fitting strategy, statistical analysis, and upper limits on the inclusive HH cross section. Furthermore, we will showcase potential improvements using Transformer-based machine learning methods compared to Boosted Decision Trees (BDT) or simple Deep Neural Networks (DNN), specifically in terms of better rejection of non-resonant backgrounds. Additionally, we will discuss the inclusion of a boosted H→bb category and estimate its impact on sensitivity for the Run 3 dataset.