Solving a Higgs detection optimization problem with quantum annealing for machine learning

The discovery of Higgs-boson decays in a background of standard-model processes was assisted by machine learning methods. The classifiers used to separate signal from background are trained using high quality but imperfect simulations of the physical processes involved, resulting in systematic errors. Here we use quantum and simulated annealing to solve this binary classification problem, mapped to the problem of finding the ground state of a corresponding Ising spin model. We build a set of weak classifiers based on the kinematic observables of the Higgs decay photons, which we then use to construct a strong classifier which is resilient against overtraining and certain systematic errors in the training set. We show that these annealing-based classifier systems perform comparably to the state-of-the-art machine learning methods that are currently used in particle physics while still being simple functions of directly interpretable experimental parameters with clear physical meaning. This technique may find application in other areas of HEP, such as real-time decision making in event-selection problems.