I would like to start this thread where the community can share teasers for JuliaCon 2021 talks. This content may be useful to community leaders (cc: @logankilpatrick) who can promote the conference in other official social media (twitter, linkedin, …).
What if we could learn functions directly on 3D meshes, which are discretizations of geospatial domains? In the example below, I am using GeoStats.jl to (1) simulate two correlated Gaussian processes on the statue of Beethoven and to (2) learn a function directly on this mesh with geostatistical learning (Hoffimann et al 2021). If you are interested in this research, come watch my JuliaCon 2021 talk next month.