CERN Accélérateur de science

 
Feynman diagrams considered for $\gamma^*N$ scattering: a) photon-gluon fusion (PGF), b) gluon radiation (QCD Compton scattering), c) Leading order process (LP).
Feynman diagrams considered for $\gamma^*N$ scattering: a) photon-gluon fusion (PGF), b) gluon radiation (QCD Compton scattering), c) Leading order process (LP).
Feynman diagrams considered for $\gamma^*N$ scattering: a) photon-gluon fusion (PGF), b) gluon radiation (QCD Compton scattering), c) Leading order process (LP).
Comparison of distributions of kinematic variables between experimental and MC high-$p_T$ deuteron data.
Comparison of distributions of kinematic variables between experimental and MC high-$p_T$ deuteron data.
Comparison of distributions of kinematic variables between experimental and MC high-$p_T$ deuteron data.
Comparison of distributions of kinematic variables between experimental and MC high-$p_T$ deuteron data.
Comparison of distributions of kinematic variables between experimental and MC high-$p_T$ deuteron data.
Comparison of distributions of kinematic variables between experimental and MC high-$p_T$ deuteron data.
Comparison of distributions of kinematic variables between experimental and MC high-$p_T$ proton data.
Comparison of distributions of kinematic variables between experimental and MC high-$p_T$ proton data.
Comparison of distributions of kinematic variables between experimental and MC high-$p_T$ proton data.
Comparison of distributions of kinematic variables between experimental and MC high-$p_T$ proton data.
Comparison of distributions of kinematic variables between experimental and MC high-$p_T$ proton data.
Comparison of distributions of kinematic variables between experimental and MC high-$p_T$ proton data.
Neural network output after training the MC sample for the proton analysis. The probability of each process for a point representing an event is proportional to the distance to the opposite side of the triangle as in the ``Mandelstam representation''. An exemplary set of probabilities assigned to a point is shown by the colour lines.
Top panels: Neural network validation. Here $P_{\text{NN}}$ is the fraction of the process given by the NN and $P_{\text{MC}}$ is the true fraction of each process from MC in a given $P_{\text{NN}}$ bin. Bottom panels: Difference $P_{\text{MC}}-P_{\text{NN}}$ per bin.
Systematic changes in the final result caused by using different MC settings. Besides the final result shown on the top, seven other results are shown that are obtained with MC samples that differ by the choice of COMPASS or default LEPTO tuning, `Parton Shower’ on or off, $F_L$ from LEPTO or from $R=\sigma_L/\sigma_T$, MSTW or CTEQ5L PDF sets, FLUKA or GHEISHA for secondary interactions. The results for deuteron data are shown in the left panel and for the proton data on the right panel.
Sivers two-hadron asymmetry extracted for Photon-Gluon fusion (PGF), QCD Compton (QCDC) and Leading Process (LP) from the COMPASS high-$p_T$ deuteron (left) and proton (right) data. The $x$ range is the RMS of the logarithmic distribution of $x$ in the MC simulation. The red bands indicate the systematic uncertainties. Note the different ordinate scale used in the third row of panels.
Sivers two-hadron asymmetry extracted for Photon-Gluon fusion (PGF), QCD Compton (QCDC) and Leading Process (LP) from the COMPASS high-$p_T$ deuteron (left) and proton (right) data. The $x$ range is the RMS of the logarithmic distribution of $x$ in the MC simulation. The red bands indicate the systematic uncertainties. Note the different ordinate scale used in the third row of panels.
Collins-like two-hadron asymmetry extracted for Photon-Gluon fusion (PGF), QCD Compton (QCDC) and Leading Process (LP) from the COMPASS high-$p_T$ deuteron (left) and proton (right) data. The $x$ range is the RMS of the logarithmic distribution of $x$ in the MC simulation. The red bands indicate the systematic uncertainties.
Collins-like two-hadron asymmetry extracted for Photon-Gluon fusion (PGF), QCD Compton (QCDC) and Leading Process (LP) from the COMPASS high-$p_T$ deuteron (left) and proton (right) data. The $x$ range is the RMS of the logarithmic distribution of $x$ in the MC simulation. The red bands indicate the systematic uncertainties.