Executive Summary : | "Cosmological structure formation is a crucial tool for understanding fundamental physics, complementing terrestrial experiments. Cosmological surveys map out the Universe's structures in detail, with the clustering and distribution of these structures influenced by fundamental physics processes. Most cosmology analyses focus on the two-point function, but non-Gaussian clustering offers more information on smaller scales. To optimize the use of vast cosmological datasets and achieve critical science objectives, summary statistics sensitive to beyond-Gaussian clustering are needed. The k-Nearest Neighbor (kNN) summary statistics have been developed, sensitive to all N-point functions of clustering, both auto and cross-correlations. These summary statistics have been shown to be highly effective in extracting more information about cosmological parameters from the same data compared to two-point functions.
The proposal aims to apply these summary statistics to three science cases: searching for primordial non-Gaussianities (PNG) in matter fluctuations in the Universe, detecting cosmological 21cm clustering signals through cross-correlations with galaxy positions, and building models to predict these statistics as a function of cosmological and galaxy formation parameters. This is essential for a full analysis of all cosmological data and improving constraints on cosmological and galaxy formation models." |