Primeiros passos na obtenção de parâmetros cosmológicos utilizando matrizes de covariância cosmológicas sem ruído

Paper highlights:

• Proposes a convolutional neural network pipeline to denoise cosmological covariance matrices estimated from limited data.
• Shows that noisy covariances built from few power spectra can be significantly improved, approaching those obtained from much larger simulation sets.
• Demonstrates that the denoised covariances enable more accurate recovery of cosmological parameters in MCMC analyses.

You can read the paper here.