The recent measurement of the anomalous magnetic moment of the muon by Fermilab points
to possible new physics and undiscovered particles. However, it also necessitates validation
from other experiments. The proposed J-PARC muon g-2/EDM experiment aims to measure the
anomalous magnetic moment of the muons with a precision of 0.01 ppm using a technique different
from that used by Fermilab. The J-PARC technique uses a low emittance muon beam injected
into a storage magnetic field, eliminating the need for electric fields for focussing. The muons
decay to positrons, the hits of which are tracked by silicon detectors placed inside a cylindrical
geometry. The reconstruction of the positron tracks play a vital role in the experiment. However,
simulation studies have indicated that the current Hough transform based approach of track finding
is time consuming under pile-up conditions and further that a 40 fold improvement is essential.
Therefore, alternate track finding approaches have been proposed and are being tested. In this, we
provide an overview and status of our attempt to develop a Graph Neural Network based model
for track finding for the J-PARC muon g-2/EDM experiment