I mainly work on the 21cm signal, detected thanks to the intensity mapping (IM) technique. Doing 21cm IM means mapping the neutral hydrogen (source of the 21cm radiation and tracer of matter structures in the universe) in the observed area of the sky, frequency by frequency i.e. in a tomographic way. It is therefore a cosmology-rich probe and indeed many instruments around the world are deploying it or planning to do so (e.g. Chime, Bingo, MeerKAT, Tianlai, Fast, Hirax and eventually the SKA).
Before coming to CosmoStat, I focused on the modelling of the signal: how the atomic hydrogen is distributed in the universe, what to expect in different dark matter and dark energy models and in general the science we can do with this signal alone or in cross-correlation with others.
Nevertheless, the 21cm radiation we would like to map out is literally buried under foregrounds which are 3-5 orders of magnitude more intense. Before starting to do any cosmology, we need to face the challenge of cleaning the intensity maps. This is the reason why I am now in CosmoStat. We are now tackling this problem, doing blind source separation of the different components contaminating our 21cm intensity maps, using sparse representations of the signals, optimising it for the 21cm IM and assessing what are the limitations on the science we are able to do with this probe.