'Hello, World!': Making GNNs Talk with LLMs
This addresses the interpretability issue in graph neural networks for researchers and practitioners, though it is incremental as it builds on existing GNN and LLM techniques.
The authors tackled the problem of GNNs being black boxes by proposing Graph Lingual Network (GLN), which uses human-readable text representations based on LLMs, enabling intuitive analysis of node representation changes and achieving strong zero-shot performance on node classification and link prediction, outperforming existing LLM-based baselines.
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.