Our universe is not only expanding, but its expansion is accelerating. The nature of the "dark energy" thought to power this acceleration is one of the great mysteries in science. The best places to study this acceleration are cosmic voids—the empty spaces in between neighborhoods of galaxies—where the pattern of structure in the universe reflects fundamental physics.
We will build a cosmic void detector to map these voids and measure their physical properties deep into the universe. This virtual detector will employ Artificial Intelligence (AI) algorithms, based on deep learning neural networks, to analyze the spatial distribution of galaxies in new massive galaxy surveys.
Using cosmological simulations, we will optimize our void detector for tests that probe dark energy. We will obtain new cosmological constraints by applying this detector to early results of the imminent DESI survey and we will prepare this detector for analyses of other upcoming surveys. We will illustrate how cosmic voids evolve and how this void detector works using high-resolution graphics and movies.
As the inventors of current void-finding methods, we expect that our void detector will be widely used for analyses of the next generation of sky surveys, from which we will obtain new constraints on the properties of dark energy and other cosmological parameters.
