Reproduce “Connections and Equivalences between the Nyström Method and Sparse Variational Gaussian Processes” on OpenResearch

Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.

Universal app for Apple silicon and Intel. Requires macOS 11 or later.

About this paper

Veit Wild, Motonobu Kanagawa, Dino Sejdinovic

The paper shows that, with fixed inducing inputs, SVGP’s optimized mean equals the Nyström kernel ridge regression (KRR) fit when the inducing-point kernel matrix is invertible and GP noise variance equals the training-set size times KRR regularization. It identifies the ELBO mean objective with KRR’s regularized least-squares objective. This is not full-posterior equivalence or a measured performance gain.

Summary by alphaXiv.

Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.