EvoPath: Evolutionary Meta-path Discovery with Large Language Models for Complex Heterogeneous Information Networks

TL;DR

EvoPath is a novel framework that uses large language models to efficiently discover high-quality meta-paths in complex heterogeneous information networks, improving reasoning capabilities.

Contribution

The paper introduces EvoPath, the first framework to mitigate LLM challenges in meta-path discovery within HINs, enhancing accuracy and efficiency.

Findings

EvoPath outperforms existing methods in HIN reasoning tasks.

Effective prompting enables LLMs to generate superior meta-paths.

Ablation studies confirm the importance of each module in EvoPath.

Abstract

Heterogeneous Information Networks (HINs) encapsulate diverse entity and relation types, with meta-paths providing essential meta-level semantics for knowledge reasoning, although their utility is constrained by discovery challenges. While Large Language Models (LLMs) offer new prospects for meta-path discovery due to their extensive knowledge encoding and efficiency, their adaptation faces challenges such as corpora bias, lexical discrepancies, and hallucination. This paper pioneers the mitigation of these challenges by presenting EvoPath, an innovative framework that leverages LLMs to efficiently identify high-quality meta-paths. EvoPath is carefully designed, with each component aimed at addressing issues that could lead to potential knowledge conflicts. With a minimal subset of HIN facts, EvoPath iteratively generates and evolves meta-paths by dynamically replaying meta-paths in the buffer with prioritization based on their scores. Comprehensive experiments on three large, complex HINs with hundreds of relations demonstrate that our framework, EvoPath, enables LLMs to generate high-quality meta-paths through effective prompting, confirming its superior performance in HIN reasoning tasks. Further ablation studies validate the effectiveness of each module within the framework.