A convolutional-neural-network estimator of CMB constraints on dark matter energy injection

TL;DR

This paper introduces a convolutional neural network approach to efficiently analyze Planck 2018 data, providing model-independent constraints on dark matter annihilation effects on the cosmic microwave background.

Contribution

The study presents a novel neural network-based method to rapidly and accurately estimate likelihoods for dark matter energy injection constraints from CMB data.

Findings

Neural network accurately predicts likelihoods for dark matter annihilation

Method provides model-independent bounds on annihilation cross-section

Approach is computationally efficient and scalable

Abstract

We show that the impact of energy injection by dark matter annihilation on the cosmic microwave background power spectra can be apprehended via a residual likelihood map. By resorting to convolutional neural networks that can fully discover the underlying pattern of the map, we propose a novel way of constraining dark matter annihilation based on the Planck 2018 data. We demonstrate that the trained neural network can efficiently predict the likelihood and accurately place bounds on the annihilation cross-section in a $\textit{model-independent}$ fashion. The machinery will be made public in the near future.