Exploring the Potential of Large Language Models (LLMs) in Learning on Graphs

TL;DR

This paper investigates how Large Language Models can enhance or replace traditional Graph Neural Networks in node classification tasks, revealing new insights and promising directions for integrating LLMs into graph learning workflows.

Contribution

It introduces two novel pipelines for applying LLMs in graph learning—LLMs-as-Enhancers and LLMs-as-Predictors—and provides comprehensive empirical analysis of their effectiveness.

Findings

LLMs can significantly improve node classification accuracy.

Using LLMs as standalone predictors shows promising results.

Insights suggest new research directions for LLMs in graph learning.

Abstract

Learning on Graphs has attracted immense attention due to its wide real-world applications. The most popular pipeline for learning on graphs with textual node attributes primarily relies on Graph Neural Networks (GNNs), and utilizes shallow text embedding as initial node representations, which has limitations in general knowledge and profound semantic understanding. In recent years, Large Language Models (LLMs) have been proven to possess extensive common knowledge and powerful semantic comprehension abilities that have revolutionized existing workflows to handle text data. In this paper, we aim to explore the potential of LLMs in graph machine learning, especially the node classification task, and investigate two possible pipelines: LLMs-as-Enhancers and LLMs-as-Predictors. The former leverages LLMs to enhance nodes' text attributes with their massive knowledge and then generate predictions through GNNs. The latter attempts to directly employ LLMs as standalone predictors. We conduct comprehensive and systematical studies on these two pipelines under various settings. From comprehensive empirical results, we make original observations and find new insights that open new possibilities and suggest promising directions to leverage LLMs for learning on graphs. Our codes and datasets are available at https://github.com/CurryTang/Graph-LLM.