Constraining Effective Field Theories with Machine Learning

Abstract

We present powerful new analysis techniques to constrain effective field theories at the LHC. By leveraging the structure of particle physics processes, we extract extra information from Monte Carlo simulations, which can be used to train neural network models that estimate the likelihood ratio. These methods scale well to processes with many observables and theory parameters, do not require any approximations of the parton shower or detector response, and can be evaluated in microseconds. We show that they allow us to put significantly stronger bounds on dimension-six operators than existing methods, demonstrating their potential to improve the precision of the LHC legacy constraints.

Benchmarking simplified template cross sections in W H production

“…Recently, however, a new technique was developed that solves this problem with a combination of matrix-element information and machine learning: the Sally algorithm introduced in Refs. [23][24][25] allows us to train a neural network to estimate t(x) as a function of the observables x. * This way the score defines optimal observables not only at the parton level, but including the effects of invisible and undetected particles, parton showers, and detector response.…”

“…In this situation, the Fisher information approximation is no longer accurate, and while an analysis based on the score will still lead to correct confidence limits, they might no longer represent the best possible limits. One way to discuss this case is to use machine-learning techniques to estimate the full likelihood or likelihood ratio function to all orders in 1/Λ 2 [23][24][25]52]. In a related way, using the geometric interpretation of the Fisher information, one could calculate distances along geodesics that capture these higher-order effects as well [21].…”

“…The Fisher information is an ideal tool to analyze questions on where in phase space we can search for new physics signals and how we can extract this information from kinematic analyses [21,22]. The first question we discuss is the impact of detector effects, which can hide information that in principle exists at the parton level but is not accessible in a realistic measurement [23,24]. As described in Sec.…”

“…Until now, a major limiting factor in this approach was that detector effects and invisible particles like neutrinos could not accurately be taken into account. This problem was recently solved with a technique that combines matrix element information and machine learning [23][24][25], automated in the MadMiner tool [26].…”

Exploring anomalous couplings in Higgs boson pair production through shape analysis

“…While the benchmark points representing a cluster are isolated points in the parameter space, the procedure we propose here allows us to associate certain shapes more straightforwardly with distinct regions in the parameter space. The application of machine learning techniques in high energy physics, in particular to constrain the EFT/new physics parameter space, has been brought forward already some time ago [68][69][70][71], with successful applications in jet and top quark identification [72][73][74][75][76][77][78][79][80][81], new physics searches [70,71,[82][83][84][85][86][87][88][89][90] and PDFs [91]. Shape analysis with machine learning has been applied already to constrain anomalous Higgs-vector boson couplings in HZ production [92].…”