Resurrecting $ b\overline{b}h $ with kinematic shapes

Christophe Grojean, Ayan Paul, Z. QianarXiv 2020 · DOI · arXiv

Abstract

The associated production of a $b\bar{b}$ pair with a Higgs boson could provide an important probe to both the size and the phase of the bottom-quark Yukawa coupling, $y_b$. However, the signal is shrouded by several background processes including the irreducible $Zh, Z\to b\bar{b}$ background. We show that the analysis of kinematic shapes provides us with a concrete prescription for separating the $y_b$-sensitive production modes from both the irreducible and the QCD-QED backgrounds using the $b\bar{b}\gamma\gamma$ final state. We draw a page from game theory and use Shapley values to make Boosted Decisions Trees interpretable in terms of kinematic measurables and provide physics insights into the variances in the kinematic shapes of the different channels that help us complete this feat. Adding interpretability to the machine learning algorithm opens up the black-box and allows us to cherry-pick only those kinematic variables that matter most in the analysis. We resurrect the hope of constraining the size and, possibly, the phase of $y_b$ using kinematic shape studies of $b\bar{b}h$ production with the full HL-LHC data and FCC-hh.

Introduction

In the Standard Model, the Higgs boson couples to fermions via Yukawa couplings. While the coupling to the bottom quark ($y_b$) has been measured through Higgs decays ($h \to b\bar{b}$), measuring it through associated production ($b\bar{b}h$) remains a significant challenge. The primary obstacle is the overwhelming background from processes like $Zh$ production, which has a cross-section orders of magnitude larger than the signal and produces nearly identical one-dimensional kinematic distributions. This paper demonstrates that "kinematic shapes"—the complex, multi-dimensional correlations between particles—hold the key to disentangling the signal from these look-alike backgrounds.

Theoretical Framework and Production Modes

The $b\bar{b}h$ final state receives contributions from several distinct physical processes. The signal of interest is mediated by the bottom Yukawa coupling ($y_b$). However, it is contaminated by backgrounds mediated by the top Yukawa coupling ($y_t$) and the electroweak $Zh$ production mode.

Feynman diagrams illustrating the primary production modes for the $b\bar{b}h$ final state, categorized by their dependence on $y_b$, $y_t$, and the $Zh$ associated production.

As shown in the diagrams above, the signal involves $t$-channel exchanges, whereas the $Zh$ background is primarily $s$-channel. These topological differences manifest as subtle but exploitable differences in the multi-dimensional kinematic correlations of the final state particles.

Methodology

To isolate the $y_b$ signal, the authors focus on the $b\bar{b}\gamma\gamma$ final state, utilizing the clean diphoton decay of the Higgs to minimize combinatorial noise. The analysis employs advanced multivariate techniques:

  1. Machine Learning Models: Boosted Decision Trees (XGBoost) and Deep Neural Networks (TensorFlow) are used to model the high-dimensional space of kinematic variables, including transverse momenta ($p_T$), invariant masses (e.g., $m_{bb}$, $m_{b1h}$), and angular correlations.
  2. Interpretability with Shapley Values (SHAP): Borrowing from cooperative game theory, SHAP values are used to identify which kinematic correlations the model prioritizes. This "opens the black box," providing physical insight into why certain variables, like the correlation between Higgs momentum and the $b$-jet system, are effective discriminators.
  3. Simulation Pipeline: Datasets were generated using MadGraph5_aMC@NLO, showered with Pythia8, and processed through Delphes detector simulations tailored for the HL-LHC (14 TeV) and FCC-hh (100 TeV) environments.

Key Results

The study reveals that while 1D histograms overlap significantly, the signal and background occupy different regions in multi-dimensional space.

  • HL-LHC Prospects: The analysis projects a statistical significance of 3.33$\sigma$ for the $y_b^2$ signal with 6 $ab^{-1}$ of data. While not a standalone discovery, it provides a vital independent constraint on the bottom Yukawa coupling.
  • FCC-hh Dominance: At the 100 TeV FCC-hh, the significance increases dramatically to 63.7$\sigma$ (statistical). Even with a conservative 1% systematic error, the significance remains a robust 17.9$\sigma$.
  • CP Violation: The $b\bar{b}h$ channel is sensitive to the phase of $y_b$ through interference terms. Including this channel in global fits improves constraints on the CP-violating phase of the bottom Yukawa coupling by approximately 15% at the FCC-hh compared to using Higgs decay measurements alone.

Conclusion

By moving beyond traditional cut-based analyses to embrace kinematic shape studies and interpretable machine learning, this research resurrects the $b\bar{b}h$ channel as a precision tool. This approach not only enhances the sensitivity of the HL-LHC but establishes $b\bar{b}h$ as a "golden" probe for the FCC-hh, capable of resolving the size and phase of the bottom-quark Yukawa coupling with unprecedented clarity.

Summary updated

Abstract

The associated production of a $b\bar{b}$ pair with a Higgs boson could provide an important probe to both the size and the phase of the bottom-quark Yukawa coupling, $y_b$. However, the signal is shrouded by several background processes including the irreducible $Zh, Z\to b\bar{b}$ background. We show that the analysis of kinematic shapes provides us with a concrete prescription for separating the $y_b$-sensitive production modes from both the irreducible and the QCD-QED backgrounds using the $b\bar{b}\gamma\gamma$ final state. We draw a page from game theory and use Shapley values to make Boosted Decisions Trees interpretable in terms of kinematic measurables and provide physics insights into the variances in the kinematic shapes of the different channels that help us complete this feat. Adding interpretability to the machine learning algorithm opens up the black-box and allows us to cherry-pick only those kinematic variables that matter most in the analysis. We resurrect the hope of constraining the size and, possibly, the phase of $y_b$ using kinematic shape studies of $b\bar{b}h$ production with the full HL-LHC data and FCC-hh.
BibTeX
@misc{grojean2020resurrecting,
  author = {Christophe Grojean and Ayan Paul and Z. Qian},
  title = {{Resurrecting $ b\textbackslash\{\}overline\{b\}h $ with kinematic shapes}},
  year = {2020},
  eprint = {2011.13945},
  archivePrefix = {arXiv},
  url = {https://arxiv.org/abs/2011.13945},
}
Figure