Semantic Refinement with LLMs for Graph Representations

ArXi:2512.21106v2 Announce Type: replace-cross Graph-structured data exhibit substantial heterogeneity in where their predictive signals originate: in some domains, node-level semantics dominate, while in others, structural patterns play a central role. This structure-semantics heterogeneity implies that no graph learning model with a fixed inductive bias can generalize optimally across diverse graph domains.