Simultaneous Estimation of Elliptic Flow Coefficient and Impact Parameter in Heavy-Ion Collisions using CNN

TL;DR

This paper introduces a CNN-based method for the simultaneous estimation of elliptic flow coefficient and impact parameter in heavy-ion collisions, demonstrating accurate predictions aligned with simulated and experimental data at LHC energies.

Contribution

It is the first to develop a CNN model that predicts both $v_{2}$ and impact parameter simultaneously in heavy-ion collision analysis.

Findings

CNN preserves centrality and $p_{T}$ dependence of $v_{2}$

Accurately predicts impact parameter with low error margins

Models trained on AMPT simulated Pb-Pb collision data

Abstract

A deep learning based method with Convolutional Neural Network (CNN) algorithm is developed for simultaneous determination of the Elliptic Flow coefficient ($v_{2}$) and the Impact Parameter in Heavy-Ion Collisions at relativistic energies. The proposed CNN is trained on Pb$-$Pb collisions at $\sqrt{s_{NN}}$ = 5.02 TeV with minimum biased events simulated with the AMPT event generator. A total of twelve models were built on different input and output combinations and their performances were evaluated. The predictions of the CNN models were compared to the estimations of the simulated and experimental data. The deep learning model seems to preserve the centrality and $p_{T}$ dependence of $v_{2}$ at the LHC energy together with predicting successfully the impact parameter with low margins of error. This is the first time a CNN is built to predict both $v_{2}$ and the impact parameter simultaneously in heavy-ion system.