The concept of data challenges is fairly new and originates from the artificial intelligence community. It is now used by many instances to address targeted problems. The principle is straightforward: a problem of interest is introduced, adequately described and a set of data is provided to the participants in order to solve it. The goal is then to find the most accurate solution according to predefined metrics. Data challenges are often proposed by companies or entities to obtain a unique solution, the best solution, to their problem.
In our case, Datlas advocates for a more collaborative data challenge where the problem is co-defined in conjunction with the participants. Also, the algorithms used by the participants are open and shared. The results of all the methods are then discussed and analyzed as a group. This way, collaborative data challenges are appropriate platforms to gather forces and unite scientific groups. Moreover, in this configuration, the outcome of the challenges is not only the best solution to a problem but also a more nuanced and rich scientific understanding.
A challenge on mapping real altimetric data in the Mediterranean Sea created by Datlas and MEOM-IGE. [...]
A challenge on mapping high frequency SSH with artificial SWOT and nadir data in the Californian SWOT X-over created by Datlas and MEOM-IGE. [...]
A challenge on the SWOT Karin instrumental error filtering organised by Datlas, IMT Altlantique and CLS. [...]
A challenge on mapping pseudo altimetric data on a QG model created by Datlas and MEOM-IGE. [...]
A challenge on the SWOT satellite error calibration organised by Datlas, IGE, IMT Altlantique and CLS. [...]
The goal is to investigate how to best reconstruct sequences of Sea Surface Height (SSH) maps from real nadir satellite altimetry observations. [...]
The goal is to investigate how to best reconstruct sequences of Sea Surface Height (SSH) maps artificial nadir and SWOT satellite altimetry observations. [...]