Speaker
Description
I present ASTRA (Algorithm for Stochastic Topological RAnking), a new method for classifying cosmic web structures, designed to explore the dark universe. While traditional approaches struggle to map both dense regions and cosmic voids—critical tracers of dark matter and cosmic acceleration—ASTRA leverages probabilistic reconstruction of underdense regions using random catalogs. This allows us to trace void cores, filaments, and sheets with unprecedented completeness, even in sparsely sampled volumes.
Using DESI Data Release 2 (DR2, not yet public), I show how ASTRA's cosmic web classifications provide high-precision constraints on large-scale structure. These maps reveal subtle spatial correlations in the distribution of voids and filaments, giving new ways to probe dark matter halo assembly and dark energy's influence on cosmic expansion. The method's computational efficiency makes it particularly powerful for next-generation surveys, enabling cosmological parameters exploration beyond standard two-point analyses.
