Reproduce “Attention Graphons: A Graph Limit Perspective on Graph Transformers” 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
Caio F. Deberaldini Netto, Moshe Eliasof, Luana Ruiz
The authors analyze whether graph transformer attention can be summarized by stable, dataset-specific graphons. They symmetrize bounded pre-softmax scores, sort nodes by degree, and average interactions in blocks. In most reported datasets, cut-norm residual variance generally decreases with graph size, consistent with stabilization; REDDIT-MULTI-5K exceeds the O(n^-1) proxy at its two largest sizes, so stability is not universal.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.