The full exploitation of science data is being impeded by statistical and computational intractability. Cosmologists are now deluged by large, complex datasets, and the scientific value of these data is presently lessened by the laborious efforts needed to analyse them. The primary goal of this project is to tackle this problem through the combination of new computer science and statistics by developing, and deploying on the Grid, efficient algorithms for the analysis of massive cosmological datasets e.g. fast, tree-based versions of the n-point correlation functions and Kernel density estimation. A secondary goal is the use of these algorithms to obtain more precise measurements of the dark energy content of the Universe, particularly as a function of red shift. Such measurements will provide critical constraints on the origin of dark energy: one of the biggest conundrums in science.
This will be achieved using these new catalogues of galaxies and quasars from the completed Sloan Digital Sky Survey. This pro ject will be done in collaboration with the US Virtual Observatory, UK AstroGrid, the department of the Institute of Cosmology and Gravitation (ICG) and the Pittsburgh Computational AstroStatistics (PiCA) group. Such multi-disciplinary research facilitates new fundamental physics and demonstrates the synergy between statistics, computer science and astrophysics. The algorithms and techniques developed during this project will also be invaluable to other areas of science and industry e.g. biology, manufacturing.
Dr. Robert Nichol is a recognized leader in the development of data-mining algorithms for massive astronomical datasets. He returns to the UK (ICG) after 12 years in the US. The ICG has offered him a permanent position, and is committed to support the project for at least 3 years. Dr Nichol will bring a new aspect of high-tech training to researchers at the ICG and throughout Europe, as well as fostering new US-European collaborations.
Fields of science
Call for proposal
See other projects for this call