Tag
This paper develops NTK-KIP, MetaQuill, and MetaQuill-KIP algorithms to improve neural field reconstruction from sparse observations, making NTK-driven neural fields non-linear and meta-learnable for efficient few-shot adaptation.