ahhh that was the piece that was missing for me. So I will want to implement multiple realizations of the same covariance matrix but pulling from a different gp_helper. There are actually two semi-hierarchical setups here - I’m assuming the length_scale and beta parameters will differ by person (which is not yet implemented in this version of the code), but each person also sees different grids that I’m assuming have the same length-scale and beta parameters (within each person) but are different realizations of the GP. I’m pretty familiar with hierarchical models outside of GPs so I think I understand this well enough to use now. I’ll try out these suggestions and report back with the (hopefully) functioning version. thanks very very much!
2 Likes