CERN Accelerating science

: justification=centerlast

: Observables reconstructed in production of the on-shell Higgs boson in the $ZZ$ and $WW$ fusion: azimuthal angle between two associated jets $\Delta\Phi_\mathrm{JJ}$ (left); observable ${\cal R}_\mathrm{opt,2}$ optimized for the quadratic term $\theta_1^2$ (middle); observable ${\cal R}_\mathrm{opt,1}$ optimized for the linear term $\theta_1$ (right). Four distributions are shown: SM ($\theta_0$, black), $CP$-odd operator ($\theta_1$, red), and 50\% mixture with positive (green) and negative (blue) relative sign of $\theta_0$ and $\theta_1$. The study is inspired by Ref.~\cite{Gritsan:2020pib}.

: Caption not extracted

: justification=centerlast

: The ROC curves showing the separation power between the processes driven by a $CP$-odd coupling $\theta_1$ and the SM $\theta_0$ (left) and between the 50\% mixture of the two processes with the positive and negative sign of interference (right). Three observables in the on-shell Higgs boson production in vector boson fusion are tested: $\Delta\Phi_\mathrm{JJ}$ (blue), optimized with matrix element (red) and machine learning (green). The study is inspired by Ref.~\cite{Gritsan:2020pib}.

justification=centerlast : Schematic representation of STXS 1.2 bins for the electroweak production of the Higgs boson in association with two jets~\cite{lhcxwg}.

justification=centerlast : The chain of MC modeling of probability density ${\cal P}(\vec{x}_\mathrm{reco}|\vec{\theta})$ as a function reconstructed observables $\vec{x}_\mathrm{reco}$ and EFT parameters $\vec{\theta}$ following Eq.~(\ref{eq:P}). An event generator produces LHE files for $N$ hypotheses $\vec\theta_i$, which are processed through full MC simulation, including parton shower and detector effects, and further re-weighted into $M$ hypotheses necessary for analytical morphing of template parameterization. The diagram is inspired by Ref.~\cite{CMS:2021nnc}.

Illustration of simulation-based inference techniques that train a neural network surrogate for the likelihood or likelihood ratio function. Figure taken from Ref.~\cite{Brehmer:2018kdj}. Here $x$ corresponds to~$\vec{x}_\mathrm{reco}$. and $z$ denotes the full MC truth record.

\small Schematic representation of the datasets and their overlapping dependences on the 34 Wilson coefficients included in the analysis of Ref.~\cite{Ellis:2020unq}

Logarithm of normalised linear dependences for electroweak measurements. The entries are normalised by dividing each one by the largest operator dependence of a given measurement, $a_{\mathrm{max}}^X$, such that the colour map depicts $\log( a_i^X /a_{\mathrm{max}}^X )$.

Logarithm of normalised linear dependences for diboson measurements, as in Fig.~\ref{fig:EWPO}.

Logarithm of normalised linear dependences for each $\Delta\phi_{jj}$ bin from $-\pi$ to $\pi$ in the differential $Zjj$ measurement. The normalisation here is with respect to the strongest linear dependence across all bins of the measurement.

Logarithm of normalised linear dependences for Higgs signal strength measurements, as in Fig.~\ref{fig:EWPO}. Dependences include effects in both production and decay.

Logarithm of normalised linear dependences for Higgs STXS measurements, as in Fig.~\ref{fig:EWPO}. Although the figure is labelled ``ATLAS'', since our analysis makes use of those specific measurements, STXS definitions, and therefore SMEFT dependences, are universal. Dependences include effects in both production and the nominal, $h\to 4\ell$ decay, except the first four. These quantify the dependences of the ratio of the other Higgs decay branching fractions to the nominal one.

\small The values of the diagonal entries of the Fisher information matrix evaluated for the dataset of the global linear (left) and quadratic (right panel) SMEFiT analysis The normalisation here is such that the sum of the entries associated to each EFT coefficient adds up to 100. For entries in the heat map larger than 10, we also indicate the corresponding numerical values.

95\% CL individual limits from electroweak measurements.

95\% CL individual limits from diboson measurements.

95\% CL individual limits from Higgs measurements.

95\% CL individual limits from Higgs STXS measurements.

Comparison of the magnitude of the 95\% CL bounds on the Wilson coefficients considered in the SMEFiT~\cite{Ethier:2021bye} analysis obtained in the global fit with those of fits to restricted datasets: a top-only, a Higgs-only, and a no-diboson fit.

Comparison of the magnitude of the 95\% CL bounds on the Wilson coefficients considered in the SMEFiT~\cite{Ethier:2021bye} analysis obtained in the global fit with those of fits to restricted datasets: a top-only, a Higgs-only, and a no-diboson fit.

Comparison of the magnitude of the 95\% CL bounds on the Wilson coefficients considered in the SMEFiT~\cite{Ethier:2021bye} analysis obtained in the global fit with those of fits to restricted datasets: a top-only, a Higgs-only, and a no-diboson fit.