Constraining primordial non-Gaussianity using Neural Networks

Abstract

We present a novel approach to estimate the value of primordial non-Gaussianity ($f_{\mathrm{NL}}$) parameter directly from the Cosmic Microwave Background (CMB) maps using a convolutional neural network (CNN). While traditional methods rely on complex statistical techniques, this study proposes a simpler approach that employs a neural network to estimate $f_{\mathrm{NL}}$. The neural network model is trained on simulated CMB maps with known $f_{\mathrm{NL}}$ in range of $[-50,50]$, and its performance is evaluated using various metrics. The results indicate that the proposed approach can accurately estimate $f_{\mathrm{NL}}$ values from CMB maps with a significant reduction in complexity compared to traditional methods. With $500$ validation data, the $f_{\mathrm{NL}}^{\mathrm{output}}$ against $f_{\mathrm{NL}}^{\mathrm{input}}$ graph can be fitted as $y=ax+b$, where $a=0.980_{-0.102}^{+0.098}$ and $b=0.277_{-0.101}^{+0.098}$, indicating the unbiasedness of the primordial non-Gaussianity estimation. The results indicate that the CNN technique can be widely applied to other cosmological parameter estimation directly from CMB images.

12 pages, 13 figures

References in corpus (8)

• Five-Year Wilkinson Microwave Anisotropy Probe (WMAP) Observations: Cosmological Interpretation
• Five-Year Wilkinson Microwave Anisotropy Probe (WMAP) Observations: Likelihoods and Parameters from the WMAP data
• Fast Automated Analysis of Strong Gravitational Lenses with Convolutional Neural Networks
• Non-Gaussianity from Preheating
• Lattice calculation of non-Gaussianity from preheating
• Recovering the CMB Signal with Machine Learning
• Likelihood-free Inference with Mixture Density Network
• Snowmass 2021 CMB-S4 White Paper