AI RESEARCH
LLM as Graph Kernel: Rethinking Message Passing on Text-Rich Graphs
arXiv CS.LG
•
ArXi:2603.14937v1 Announce Type: new Text-rich graphs, which integrate complex structural dependencies with abundant textual information, are ubiquitous yet remain challenging for existing learning paradigms. Conventional methods and even LLM-hybrids compress rich text into static embeddings or summaries before structural reasoning, creating an information bottleneck and detaching updates from the raw content. We argue that in text-rich graphs, the text is not merely a node attribute but the primary medium through which structural relationships are manifested. We.