April-June 2020

Organized by Jean-Luc Starck (CEA/DAp Saclay) and Valeria Pettorino (CEA/DAp Saclay)

I. Rationale

The nature of both dark matter and dark energy, which on their own account for 95% of the content of our Universe, remains one of the biggest mysteries in fundamental physics, which could be solved in the next decade. Indeed, the decade 2020-2030 will be cosmology’s golden age with the launch of two spatial missions dedicated to cosmology, Euclid in 2022 and WFIRST in 2026, and the starting of the Large Synoptic Survey Telescope (LSST) observations in 2022. These international projects, ushering in the era of Big Data for cosmology, will require new statistical methods to tackle computational and theoretical challenges. The GOLD program will create interactions between scientists of different discipines: new results in applied mathematics or machine learning would certainly bring in new ideas to analyse cosmological data. The GOLD program will be a highway for transferring the knowledge from statistics and machine learning fields to cosmology.

II. Goals

The start of those large observations provides a great window of opportunity to organize a long-term educational program in cosmology at Institut Pascal. Our GOLD proposal has the following goals:

Prepare a new generation of researchers to analyze the huge flow of upcoming cosmological data. This requires the transfer of knowledge in both statistical methods and cosmology.

Present and discuss the state of the art results in different areas of cosmology: the field is evolving very quickly, and having a set of dedicated workshops will help to follow and understand the latest results, and interact with the key players in the field.

Create interactions between scientists of different disciplines, new results in applied mathematics and machine learning will certainly bring new ideas to analyze cosmological data. Such a program would be a highway for transferring the knowledge from statistics and machine learning fields to the field of cosmology.

Promote reproducible research: With enormous volumes of data and highly complex algorithms, it is often impossible for a researcher to reproduce the figures published in an article.  However, the reproducibility of results lies at the very heart of the scientific approach and constitutes one of the major problems of modern science. Hence the principle which consists in publishing the source codes operated to analyse the data and the scripts used to process the data and generate the figures, in addition to the results.