I'm curious how this relates to MDL-tied regularization. I don't see any explicit reference to the L2 norm here, and I'm not too smart about graph neural networ

I'm curious how this relates to MDL-tied regularization. I don't see any explicit reference to the L2 norm here, and I'm not too smart about graph neural networks. But it seems like overfitting is almost necessary unless some kind of obscure variational technique is used? They did have one paper where they used a progressively compressed GNN to isolate an approximation for a cosmological equation smaller and more accurate than had been previously achieved. I'll update here if I find it. Edit: Ah, here we are: https://arxiv.org/abs/2006.11287 https://arxiv.org/abs/2006.11287 I believe there was some very good followup work posted on it this year that I saw some movement during the 3-5 months that I was really active on Twitter (brief time that it was), seems like an extremely promising area, and one I should learn if I truly do care about the Kolmogorov complexity of a problem.