Neural networks improve maps of cosmic matter using galaxy images
Neural Posterior Estimation for Tomographic Weak Lensing Mass Mapping
Computer Vision and Pattern Recognition
Summary
Mapping how matter, including invisible dark matter, is spread across the universe is tricky because it involves interpreting subtle distortions in images of galaxies caused by gravity. The usual methods use complicated steps and lots of computing power to estimate these distortions and rebuild maps of matter. The authors suggest using a type of AI called a neural network to directly learn from galaxy images and quickly predict the distribution of matter, including both normal and dark matter, in space. Their approach works well on simulated data, producing reliable estimates and making it easier to study the universe’s large-scale structure and its evolution.
weak gravitational lensingshearconvergencedark matterdeep neural networkvariational distributionsimulation-based inferencecosmic structureLSSTposterior estimation
Authors
Tim White, Shreyas Chandrashekaran, Camille Avestruz, Jeffrey Regier, the LSST Dark Energy Science Collaboration
Abstract
Weak gravitational lensing shear and convergence trace the distribution of baryonic and dark matter across space, making them a powerful probe of cosmic structure. Inferring shear and convergence from images is a challenging inverse problem. The prevailing approach to this task estimates shear from weighted averages of galaxy ellipticities, calibrates these estimates to account for systematic biases, and transforms them to reconstruct convergence, a multistage procedure that requires substantial computational resources and meticulous handling of statistical uncertainties. As an alternative, we propose a probabilistic approach to field-level weak lensing inference in which we train a deep neural network to directly map a multiband image to a variational distribution over the underlying tomographic shear and convergence fields. This neural posterior estimation (NPE) procedure implicitly marginalizes over nuisance variables in the cosmological forward model and does not require evaluating the likelihood function. It is also amortized, so it enables rapid posterior inference for astronomical surveys once the neural network is trained. When evaluated on synthetic images from the LSST-DESC DC2 Simulated Sky Survey, NPE produces well-calibrated variational distributions for shear and convergence that are consistent with the ground truth. We describe how maps sampled from these variational distributions could be used in a subsequent simulation-based inference procedure to approximate the posterior distribution over cosmological parameters.