Reproduce “Graph Neural Tangent Kernel: Convergence on Large Graphs” 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
Sanjukta Krishnagopal, Luana Ruiz
The authors study whether a graph neural tangent kernel fitted on a smaller graph can transfer to a larger graph for node-level prediction. They prove convergence of these kernels and their spectra under dense-graph assumptions, and test transfer on several tasks. On MovieLens-100k, relative test-MSE difference was 0.09% (standard deviation 0.03 percentage points); this compares errors, not accuracy or computation.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.