Reproduce “ReCoG: Relative Node Contexts for Multi-Domain Graph Pre-training” 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
Yupeng Tong, Lianze Shan, Jitao Zhao, Dongxiao He
ReCoG transfers graph representations by constructing attribute-, structural-, and semantic contexts with shared ranking rules, while retaining information from raw links. It pre-trains an attribute-reconstruction model and adapts its frozen encoder using target supervision. In one-shot node classification across six targets, it averages 63.08% versus 55.47% for TIG, ranking first on five targets; GraphLoRA leads on PubMed.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.