Speaker
Description
Abstract: In recent years, deep learning has found many applications in the field of
high energy heavy-ion collisions. Deep learning technique provides a data-driven statistics based model which can help map the input and output observables. In such cases
the mapping function is difficult to formulate or is usually unknown. In this work, we
explore the prospects of using deep learning methods such as a feed-forward deep neural
network (DNN) to estimate the elliptic flow (v2) in heavy-ion collisions at the RHIC and
LHC energies. A novel method is developed to process the input observables directly
from the particle kinematic information. The proposed DNN model is trained with Pb–
Pb collisions at √
sNN = 5.02 TeV minimum bias events simulated with AMPT model.
We proceed to show that the DNN learns and preserves the centrality, energy, transverse
momentum dependence of v2 very well. We extend this work further to estimate v2 for
light-flavour hadrons such as pions, kaons, and protons. The baryon-meson v2, and the
number of constituent quark scaling are preserved by the DNN model. Error estimation
was subjected to an event simulation with additional random noise, and the proposed
DNN model is found to keep the robustness and prediction accuracy intact up to a reasonable extent. Results are compared to experimental data wherever possible.
References:
1. N. Mallick, S. Prasad, A. N. Mishra, R. Sahoo, and G. G. Barnaf¨oldi, Phys.Rev.D
105, 114022 (2022).
2. N. Mallick, S. Prasad, A. N. Mishra, R. Sahoo, and G. G. Barnaf¨oldi, Phys.Rev.D
107, 094001 (2023).
