Choose timezone
Your profile timezone:
Please bring a numerical question if you have one. Note we're in the small conference room now.
The one topic ended up being Alex presenting on the problem of interpolation of a smooth function to (possibly noisy) scalar function values provided at a large number of arbitrary points on the sphere. The particular smooth function class discussed was spherical harmonics up to some degree. Various methods to solve the least-squares linear system were presented, and connections drawn to NUFFTs and Gaussian process regression.