Search for Mono-Higgs Signals in $b\bar b$ Final States Using Deep Neural Networks

TL;DR

This paper introduces a Hybrid Deep Neural Network approach combining MLP and CNN architectures to improve mono-Higgs signal detection in $b\bar{b}$ final states at the HL-LHC by leveraging kinematic and color flow features.

Contribution

It presents a novel Hybrid Deep Neural Network architecture that integrates multiple data types for enhanced signal-background discrimination in mono-Higgs searches.

Findings

Hybrid DNN outperforms single-input networks in classification accuracy.

Embedding color flow structures into images improves background rejection.

Method is optimized for the High-Luminosity LHC conditions.

Abstract

We study mono-Higgs signatures emerging in an illustrative new physics scenario involving Standard Model Higgs boson decays to bottom quark pairs using Hybrid Deep Neural Networks. We use a Multi-Layer Perceptron to analyze the kinematic observables and optimize the signal-to-background discrimination. The global color flow structure of hard jets emerging from the decay of heavy particles with different color charges is crucial to single out the mono-Higgs signature. Upon embedding the different color flow structures for signal and backgrounds into constructed images, we use a Convolution Neural Network to analyze the latter. Specifically, the approach takes initially a mono-type data as input, frittering away invaluable multi-source and multi-scale information. We then discuss a general architecture of Hybrid Deep Neural Networks that supports instead mixed input data. In comparison with single input Deep Neural Networks, like MultiLayers Perceptron or Convolution Neural Network, the Hybrid Deep Neural Networks provide higher capacity in feature extraction and thus in signal vs background classification performance. We provide reference results for the case of the High-Luminosity Large Hadron Collider.