Bridging Graph Structure and Knowledge-Guided Editing for Interpretable Temporal Knowledge Graph Reasoning

TL;DR

This paper introduces IGETR, a hybrid framework combining GNNs and LLMs for interpretable and accurate temporal knowledge graph reasoning, addressing structural and semantic challenges in dynamic data.

Contribution

IGETR is a novel three-stage reasoning framework that integrates graph-based evidence, LLM-guided path editing, and path integration for improved temporal knowledge graph inference.

Findings

Achieves state-of-the-art results on ICEWS datasets.

Outperforms baselines with up to 5.6% improvement in Hits@1.

Ablation studies validate each component's effectiveness.

Abstract

Temporal knowledge graph reasoning (TKGR) aims to predict future events by inferring missing entities with dynamic knowledge structures. Existing LLM-based reasoning methods prioritize contextual over structural relations, struggling to extract relevant subgraphs from dynamic graphs. This limits structural information understanding, leading to unstructured, hallucination-prone inferences especially with temporal inconsistencies. To address this problem, we propose IGETR (Integration of Graph and Editing-enhanced Temporal Reasoning), a hybrid reasoning framework that combines the structured temporal modeling capabilities of Graph Neural Networks (GNNs) with the contextual understanding of LLMs. IGETR operates through a three-stage pipeline. The first stage aims to ground the reasoning process in the actual data by identifying structurally and temporally coherent candidate paths through a temporal GNN, ensuring that inference starts from reliable graph-based evidence. The second stage introduces LLM-guided path editing to address logical and semantic inconsistencies, leveraging external knowledge to refine and enhance the initial paths. The final stage focuses on integrating the refined reasoning paths to produce predictions that are both accurate and interpretable. Experiments on standard TKG benchmarks show that IGETR achieves state-of-the-art performance, outperforming strong baselines with relative improvements of up to 5.6% on Hits@1 and 8.1% on Hits@3 on the challenging ICEWS datasets. Additionally, we execute ablation studies and additional analyses confirm the effectiveness of each component.