'Hello, World!': Making GNNs Talk with LLMs

TL;DR

This paper introduces Graph Lingual Network (GLN), a GNN built on large language models that produces human-readable text representations, enabling interpretability and strong zero-shot performance on graph tasks.

Contribution

The paper presents a novel GNN architecture that uses LLMs for interpretable, text-based node representations, integrating advanced GNN techniques and demonstrating zero-shot capabilities.

Findings

GLN achieves strong zero-shot performance on node classification.

GLN provides human-readable, interpretable node representations.

Incorporates advanced GNN techniques like attention and residual connections.

Abstract

While graph neural networks (GNNs) have shown remarkable performance across diverse graph-related tasks, their high-dimensional hidden representations render them black boxes. In this work, we propose Graph Lingual Network (GLN), a GNN built on large language models (LLMs), with hidden representations in the form of human-readable text. Through careful prompt design, GLN incorporates not only the message passing module of GNNs but also advanced GNN techniques, including graph attention and initial residual connection. The comprehensibility of GLN's hidden representations enables an intuitive analysis of how node representations change (1) across layers and (2) under advanced GNN techniques, shedding light on the inner workings of GNNs. Furthermore, we demonstrate that GLN achieves strong zero-shot performance on node classification and link prediction, outperforming existing LLM-based baseline methods.