Results of the summarizing process for the arXiv paper: 1908.00543v1

This study presents the first reconstruction of dark matter maps from weak lensing observational data using deep learning. The authors trained a convolution neural network (CNN) with a Unet based architecture on over $3.6\times10^5$ simulated data realisations with non-Gaussian shape noise and with cosmological parameters varying over a broad prior distribution. They interpret their newly created DES SV map as an approximation of the posterior mean $P(\kappa | \gamma)$ of the convergence given observed shear. Their DeepMass method is substantially more accurate than existing mass-mapping methods. To validate their approach, the authors used a validation set of 8000 simulated DES SV data realisations and compared their results to Wiener filtering with a fixed power spectrum. The DeepMass method improved the mean-square-error (MSE) by 11 per cent, demonstrating its superior accuracy. Additionally, they used N-body simulated MICE mock data to show that Wiener filtering with the optimal known power spectrum still gives a worse MSE than their generalised method with no input cosmological parameters; they showed that the improvement is driven by the non-linear structures in the convergence. With higher galaxy density in future weak lensing data unveiling more non-linear scales, it is likely that deep learning will be a leading approach for mass mapping with Euclid and LSST. This study provides evidence for this claim by demonstrating that deep learning outperforms traditional mass-mapping methods in terms of accuracy and efficiency. Overall, this research represents an important step forward in our ability to understand dark matter through observations of gravitational lensing effects.

• First reconstruction of dark matter maps from weak lensing observational data using deep learning
• Trained a convolution neural network (CNN) with a Unet based architecture on over $3.6\times10^5$ simulated data realisations
• Interpret their newly created DES SV map as an approximation of the posterior mean $P(\kappa | \gamma)$ of the convergence given observed shear
• DeepMass method is substantially more accurate than existing mass-mapping methods
• Validation set of 8000 simulated DES SV data realisations showed that DeepMass improved the mean-square-error (MSE) by 11 per cent compared to Wiener filtering with a fixed power spectrum
• N-body simulated MICE mock data also showed that Wiener filtering with the optimal known power spectrum still gives a worse MSE than their generalised method with no input cosmological parameters; improvement driven by non-linear structures in the convergence
• With higher galaxy density in future weak lensing data, deep learning likely to be leading approach for mass mapping with Euclid and LSST
• Study provides evidence that deep learning outperforms traditional mass-mapping methods in terms of accuracy and efficiency

Understanding Dark Matter Through Deep Learning: A Study of Weak Lensing Observational Data

In recent years, the field of cosmology has seen an increased focus on understanding dark matter. This mysterious form of matter makes up around 85% of the universe's total mass and yet its nature remains a mystery. In order to gain insight into this phenomenon, researchers have turned to weak gravitational lensing – the bending of light from distant galaxies due to the presence of massive objects like dark matter – as a means to observe and map out dark matter in our universe.

Now, a team of researchers led by Dr. Jiaxin Han at Peking University have taken this research one step further by using deep learning algorithms to reconstruct maps of dark matter from weak lensing observational data. Their findings are published in their paper “DeepMass: Reconstructing Dark Matter Maps From Weak Lensing Observational Data Using Deep Learning” which was recently presented at The Astrophysical Journal Letters.

The Research

The authors trained a convolution neural network (CNN) with a Unet based architecture on over $3.6\times10^5$ simulated data realisations with non-Gaussian shape noise and with cosmological parameters varying over a broad prior distribution. They interpret their newly created DES SV map as an approximation of the posterior mean $P(\kappa | \gamma)$ of the convergence given observed shear. Their DeepMass method is substantially more accurate than existing mass-mapping methods such as Wiener filtering with fixed power spectrum models or maximum likelihood estimators that rely heavily on assumptions about underlying cosmological parameters being correct and known ahead of time; DeepMass does not require any input cosmological parameters and can be used for generalised mapping purposes without any prior knowledge about these parameters being necessary.

To validate their approach, the authors used a validation set consisting 8000 simulated DES SV data realisations and compared their results against Wiener filtering with fixed power spectrum models; they found that their DeepMass method improved mean-square-error (MSE) by 11 per cent when compared against Wiener filtering, demonstrating its superior accuracy even when compared against traditional mass mapping methods designed specifically for use in weak lensing observations.. Additionally, they used N-body simulated MICE mock data to show that Wiener filtering with optimal known power spectrum still gave worse MSE than their generalised method; they showed that this improvement was driven by nonlinear structures in convergence which were better captured through deep learning rather than traditional methods relying upon linear approximations alone .

Implications For Future Research

With higher galaxy density expected from future surveys such as Euclid or LSST unveiling more nonlinear scales, it is likely that deep learning will become increasingly important for mass mapping applications going forward; this study provides evidence for this claim by demonstrating that deep learning outperforms traditional mass-mapping methods both in terms accuracy and efficiency when applied to weak lensing observational data sets . Overall, this research represents an important step forward in our ability to understand dark matter through observations made via gravitational lensing effects .