This third challenge, involving 33 teams from 16 different countries, aimed to analyze a vast dataset simulating one of the key objectives of the future SKA (Square Kilometre Array) radio telescope : the observation of the Epoch of Reionization, a key phase in the evolution of the Universe when neutral gas was ionized by the first stars and galaxies.
The challenge was to extract the very weak signal of neutral hydrogen (the line at 21 cm) from the much more intense emissions of astronomical objects in the foreground, a bit like trying to perceive a whisper in the middle of a noisy concert.
To meet this challenge, the winning DOTSS-21 team used a combination of innovative techniques, including a machine-learning method (ML-GPR) developed as part of the LOFAR-EoR and NenuFAR Cosmic Dawn collaborations. This method accurately extracted the cosmological signal from hydrogen, despite the presence of much brighter stray emissions.
Relying on a powerful computing cluster, the team was able to process 7.5 TB of data to extract the fluctuations of this signal, simulating the Universe as it was billions of years ago.
This success opens up exciting prospects for the analysis of future SKA data, the first observations of which are expected around 2030.
It demonstrates that the methods developed by these teams are ready to meet the challenges of exploiting SKA data and isolating the 21 cm signal from Reionization. Hopes are high for such a detection, which would be a major breakthrough in astrophysics, enabling us to trace the origins of the Universe’s earliest structures and understand the processes that led to its reionization.
CNRS laboratory involved
- Laboratoire d’étude du rayonnement et de la matière en astrophysique et atmosphères (LERMA - Observatoire de Paris - PSL)
- Tutelles : CNRS / CY Cergy Paris Univ / Observatoire de Paris - PSL / Sorbonne Univ