Many popular theories of the early Universe predict that gravitational waves were generated in the very first moments of its life. These theories generally invoke new physics from beyond the standard model of the particle physics and the primordial gravitational waves are believed to carry clues about the nature of those new physics laws. Detecting the primordial gravitational waves would thus have revolutionary impact on our understanding of cosmology and fundamental physics. Currently the most promising, if not the only, way to achieving this is through observations of Cosmic Microwave Background (CMB) polarization. The taletelling bits of information are contained in a pattern of minuscule CMB polarization anisotropies, which have to be recovered from huge volumes of data collected by current and forthcoming CMB experiments and characterized with unprecedented precision and robustness. These tasks continue posing new and exciting challenges, which have to be addressed if data sets of the forthcoming, unprecedented in sensitivity and robustness, CMB observatories are to be fully analyzed and thoroughly exploited.  Addressing these challenges is the major objective of this project. This, if successful, will have huge impact on modern cosmology and our understanding of the Universe. Succeeding here will require reaching beyond the standard tools of CMB data analysts and will need to capitalize on the latest developments in data science and numerical algorithmics combined with understanding of the physics and cosmology. The project proposed here is therefore inherently interdisciplinary and will involve research in all these science areas permitting the student to develop expertise in cosmology and physics accompanied by a broad, robust background in statistics, numerical methods, high performance scientific computing and data science.

CMB polarization is also expected to shed a new light on some other important and intriguing questions of modern science, such as those concerning the nature of dark matter, total mass of neutrinos and their mass hierarchy, or the presence of unknown relativistic particle species, and which we will also target in this project.

The core of the envisaged work will involve a development of new techniques, numerical and statistical, suitable for analysis of the future CMB data, their implementation within data analysis pipelines of current and forthcoming experiments and their application to their analysis as well as a development, design and optimisation of the future experiments and experimental concepts. These will be validated and demonstrated on actual and simulated CMB data sets. The student will be based at APC and the work will be conducted within the context of an interdisciplinary ANR project, B3Dcmb, which brings together statisticians, computer scientists and cosmologists from ENSAE, INRIA and APC. They provide a stimulating, multi-disciplinary environment within which to devise novel, robust solutions to the CMB data analysis problems. The proposed PhD project will have a unique chance to capitalise on their diverse expertise and experience. In addition, 2 more PhD students, one in statistics and the other —  in high performance scientific computing, will work in parallel focusing on statistical and numerical aspects of the CMB problems. They will be based at ENSAE (Palaiseau) and INRIA (Paris), respectively, but will spend part of their time at APC on a regular basis. At APC the student will work within a close-knit, very active research group composed of two researchers, software engineer, and two current PhD students. Two postdoctoral researchers are also to be recruited within the next 3 years. The student will also benefit from a presence of a larger cosmology group, involving experts from diverse areas of cosmology.

The proposed algorithmic work will be driven and informed by the needs of cutting edge data sets as already collected by current and expected of forthcoming experiments. We will focus specifically on data sets of the currently operating ground-based POLARBEAR/Simons Array experiment, on the forthcoming ground-based Simons Observatory and a satellite mission, LiteBIRD. The student will have opportunity to become a member of these teams and take active part in their on-going efforts. In particular, s/he will be able to join the POLARBEAR/Simons Array team and contribute to, and later to help coordinate, the effort aiming at the analysis and scientific exploitation of its data sets. This work will involve direct interactions with the POLARBEAR/Simons Array team and visits to Berkeley and San Diego and will provide invaluable hands-on experience in the analysis and mining of very large data sets, but also a chance to participate in the cutting-edge science as expected of this experiments. POLARBEAR/Simons Array is one of the leading, currently operating CMB B-mode polarization experiments, which is to be upgraded in early 2018 to feature up to three telescopes observing the sky in as many as 4 frequency bands, and thus becoming one of the most advanced CMB observatories operating on this time scales.

The student will have an opportunity to be involved in Simons Observatory, a pathfinder for the future generation of new CMB observatories, expected to pave the way to the ultimate, ground-based CMB observatory. The experiment is under construction and is expected to become operational at the end of the proposed thesis work (2021). The on-going work involves an analysis pipeline development and the definition and optimisation of its instruments and operations. Alternately, s/he could apply the novel techniques in the context of the development of LiteBIRD, a new, Japan-led, CMB satellite mission, which is scheduled to be launched in 2026. The analysis methods and techniques developed as part of the proposed project are expected to become a backbone of its data analysis pipeline.

This thesis is expected to be preceded by a Master 2 internship at APC Laboratory.

The candidate should have strong background in physics and basic programming skills. Experience in numerical algorithms and knowledge of advanced programming techniques would be a plus, as would familiarity with statistical methods. However, any of these can be also acquired as part of the project.

Contact:

R.Stompor (radekapc.univ-paris-diderot.fr)
J.Errard   (josquinapc.in2p3.fr)