From Nodes to Narratives: Explaining Graph Neural Networks with LLMs and Graph Context

2025-08-09

AI Summary

Graph Neural Networks (GNNs) exhibit limited interpretability on text-augmented graphs; existing methods struggle to generate faithful, fine-grained, and semantically coherent natural-language explanations. Method: We propose LOGIC, the first framework that aligns GNN node embeddings into the latent space of a large language model (LLM) via contrastive learning, and employs hybrid soft prompting to jointly encode graph structure and textual inputs—enabling the LLM to directly generate narrative explanations grounded in GNN internal representations, while simultaneously extracting a concise explanation subgraph. Contribution/Results: LOGIC requires no post-training or human annotation, enabling end-to-end interpretable reasoning. Experiments on four real-world datasets demonstrate that LOGIC achieves superior trade-offs between explanation fidelity and sparsity, significantly enhancing human comprehensibility and analytical insight over state-of-the-art baselines.

Technology Category

• Machine Learning: Graph-based Machine Learning
• Knowledge Representation and Reasoning: Logic Programming
• Natural Language Processing: (Large) Language Models

Application Category

• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking

Abstract

Graph Neural Networks (GNNs) have emerged as powerful tools for learning over structured data, including text-attributed graphs, which are common in domains such as citation networks, social platforms, and knowledge graphs. GNNs are not inherently interpretable and thus, many explanation methods have been proposed. However, existing explanation methods often struggle to generate interpretable, fine-grained rationales, especially when node attributes include rich natural language. In this work, we introduce LOGIC, a lightweight, post-hoc framework that uses large language models (LLMs) to generate faithful and interpretable explanations for GNN predictions. LOGIC projects GNN node embeddings into the LLM embedding space and constructs hybrid prompts that interleave soft prompts with textual inputs from the graph structure. This enables the LLM to reason about GNN internal representations and produce natural language explanations along with concise explanation subgraphs. Our experiments across four real-world TAG datasets demonstrate that LOGIC achieves a favorable trade-off between fidelity and sparsity, while significantly improving human-centric metrics such as insightfulness. LOGIC sets a new direction for LLM-based explainability in graph learning by aligning GNN internals with human reasoning.

Problem

• Generating interpretable explanations for GNN predictions
• Handling rich natural language in node attributes
• Aligning GNN internal representations with human reasoning

Innovation

• LLM-based post-hoc GNN explanation framework
• Hybrid prompts combining soft and textual inputs
• Aligns GNN internals with human reasoning