In search of an interaction in the dark sector through Gaussian Process and ANN approaches

TL;DR

This study investigates potential interactions in the dark sector using Gaussian Process and Artificial Neural Networks, revealing that deviations from a cosmological constant equation of state suggest possible dark sector interactions.

Contribution

It introduces the application of ANN alongside GP for reconstructing dark sector interactions, demonstrating ANN's efficiency in this context.

Findings

Interaction is not prominent at w_DE = -1.

Evidence of interaction emerges when w_DE deviates from -1.

ANN outperforms GP in reconstructing dark sector interactions.

Abstract

Whether the current observational data indicate any evidence of interaction between the dark sector is a matter of supreme interest at the present moment. This article searched for an interaction in the dark sector between a pressure-less dark matter and a dark energy fluid with constant equation of state, $w_{\rm DE}$. For this purpose, two non-parametric approaches, namely, the Gaussian Process (GP) and the Artificial Neural Networks (ANN) have been employed and using the Hubble data from Cosmic Chronometers (CC), Pantheon+ from Supernovae Type Ia and their combination we have reconstructed the interaction function. We find that for $w_{\rm DE} =-1$, the interaction in the dark sector is not prominent while for $w_{\rm DE} \neq -1$, evidence of interaction is found depending on the value of $w_{\rm DE}$. In particularly, we find that if we start deviating from $w_{\rm DE} = -1$ either in the quintessence ($w_{\rm DE} > -1$) or phantom ($w_{\rm DE} < -1$) direction, an emergence of dark interaction is observed from both GP and ANN reconstructions. We further note that ANN which is applied for the first time in this context seems to play a very efficient role compared to GP.