CLiENT: A new tool for emulating cosmological likelihoods using deep neural networks

TL;DR

CLiENT introduces a neural network-based emulator that directly approximates the likelihood function in cosmology, enabling faster and fully differentiable likelihood evaluations for parameter inference.

Contribution

The paper presents CLiENT, a novel neural network approach that directly emulates the likelihood function, offering advantages over traditional observable emulators in speed and differentiability.

Findings

Achieves credible intervals within 0.1 sigma of true likelihood

Requires fewer than 20,000 function evaluations

Provides a surrogate likelihood with Delta chi-squared less than 0.5

Abstract

Cosmological emulation of observables such as the Cosmic Microwave Background (CMB) spectra and matter power spectra have become increasingly common in recent years because of the potential for saving computation time in connection with cosmological parameter inference or model comparison. In this paper we present CLiENT (Cosmological Likelihood Emulator using Neural networks with TensorFlow), a new method which circumvents the computation of observables in favour of directly emulating the likelihood function for a data set given a model parameter vector. We find that the method is competitive with observable emulators in terms of the required number of function evaluations, but has the distinct advantage of producing a surrogate likelihood which is completely auto-differentiable. Using less than $2 \times 10^4$ function evaluations CLiENT typically achieves credible intervals within better than $0.1 \sigma$ of those obtained using the true likelihood and single-point emulator precision better than $\Delta \chi^2 \sim 0.5$ across relevant regions in parameter space.