Angular Power Spectrum and Elliptic Flow from Event Maps in Heavy Ion Collisions

Azimuthal and polar angle distributions of particles produced in heavy ion collisions carry important information about the evolution of the Quark Gluon Plasma. The pixelization code HEALPix was created to analyze the cosmic microwave background, but its two-dimensional representation of a sphere containing pixels of equal area has a broader application to heavy ion collisions. The angular power spectrum, an application of HEALPix, is directly related to the particle flow and is calculated through the angles θ and φ on the sphere. The elliptic flow is calculated from simulated AMPT events and publicly available CMS data. Beyond the angular power spectrum, the two-dimensional pixelization may also be used with the application of convolutional neural networks (CNN) to differentiate between the presence and absence of elliptic flow. The CNN is trained using two-dimensional pixelated images from HEALPix and images of p_T and φ density from AMPT with both string melting and the default setup. We show that HEALPix with CNN can  be used to detect the presence of elliptic flow for lead-lead collisions and thus has broader applications in the study of QGP in heavy ion collisions.