Learning New Physics from an Imperfect Machine

TL;DR

This paper introduces a neural network-based method for new physics searches that effectively manages uncertainties in Standard Model predictions, enhancing hypothesis testing in high-energy physics experiments.

Contribution

It develops a novel approach integrating neural networks with maximum likelihood ratio techniques to handle uncertainties as nuisance parameters.

Findings

Effective uncertainty management demonstrated on toy problems

Implementation in multivariate setups at the LHC

Improved hypothesis testing performance

Abstract

We show how to deal with uncertainties on the Standard Model predictions in an agnostic new physics search strategy that exploits artificial neural networks. Our approach builds directly on the specific Maximum Likelihood ratio treatment of uncertainties as nuisance parameters for hypothesis testing that is routinely employed in high-energy physics. After presenting the conceptual foundations of our method, we first illustrate all aspects of its implementation and extensively study its performances on a toy one-dimensional problem. We then show how to implement it in a multivariate setup by studying the impact of two typical sources of experimental uncertainties in two-body final states at the LHC.