The primary goal of the India-based Neutrino Observatory (INO) is to establish an underground
laboratory to study neutrinos, with a specific emphasis on atmospheric neutrinos. The focus is
on obtaining precise measurements of neutrino oscillation parameters through the utilization of a
50-kiloton magnetized Iron CALorimeter (ICAL) detector. This detector has been designed to be
particularly sensitive to muon neutrinos, whose energy is determined by reconstructing the energy
of secondary particles resulting from charged current interactions between neutrinos and iron
plates. In these interactions, muons typically carry away most of the neutrino energy. In order to
accurately measure the energy of neutrinos and the oscillation parameters, a precise reconstruction
of the muon energy is also essential. Currently, muon energy reconstruction within ICAL depends
on a Kalman filter algorithm. This paper introduces an alternative approach based on a Graph
Neural Network (GNN) for energy reconstruction. The proposed GNN method aims to evaluate
its effectiveness, accuracy, and computational speed in comparison to the traditional technique.
The analysis uses the data generated through Geant4 simulations of muon energy spanning from
1 GeV to 10 GeV.