arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2310.04668cs.LGcs.AIcs.CL
收藏

基于大语言模型(LLMs)的图上无标签节点分类

Label-free Node Classification on Graphs with Large Language Models (LLMS)

Michigan State University(密歇根州立大学);Emory University(埃默里大学)

Zhikai Chen, Haitao Mao, Hongzhi Wen, Haoyu Han, Wei Jin, Haiyang Zhang, Hui Liu, Jiliang Tang

AI总结:

本文提出 LLM-GNN 流程,通过主动选择少量节点由大语言模型标注并训练图神经网络,以不到 1 美元成本在 products 数据集上实现 74.9% 的无标签节点分类准确率。

摘要:

近年来,图神经网络(GNNs)在节点分类方面取得了显著进展。然而,它们需要大量高质量标签才能保证良好的性能。相比之下,大语言模型(LLMs)在文本属性图上展现出令人印象深刻的零样本能力。但是,它们在高效处理结构数据方面面临挑战,并且推理成本较高。鉴于这些观察,本文提出了一种基于大语言模型的图上无标签节点分类流程,称为 LLM-GNN。它融合了 GNN 和 LLM 的优势,同时缓解了它们的局限性。具体而言,利用 LLM 标注一小部分节点,然后在 LLM 的标注结果上训练 GNN,以对剩余的大部分节点进行预测。LLM-GNN 的实现面临一个独特挑战:如何主动选择节点供 LLM 标注,从而提升 GNN 训练效果?如何利用 LLM 获得高质量、具有代表性和多样性的标注,从而以更低的成本提升 GNN 性能?为应对这一挑战,我们提出了一种标注质量启发式方法,并利用 LLM 得出的置信度分数来改进节点选择。全面的实验结果验证了 LLM-GNN 的有效性。特别是,LLM-GNN 在超大规模数据集 products 上以不到 1 美元的成本实现了 74.9% 的准确率。

英文摘要:

In recent years, there have been remarkable advancements in node classification achieved by Graph Neural Networks (GNNs). However, they necessitate abundant high-quality labels to ensure promising performance. In contrast, Large Language Models (LLMs) exhibit impressive zero-shot proficiency on text-attributed graphs. Yet, they face challenges in efficiently processing structural data and suffer from high inference costs. In light of these observations, this work introduces a label-free node classification on graphs with LLMs pipeline, LLM-GNN. It amalgamates the strengths of both GNNs and LLMs while mitigating their limitations. Specifically, LLMs are leveraged to annotate a small portion of nodes and then GNNs are trained on LLMs' annotations to make predictions for the remaining large portion of nodes. The implementation of LLM-GNN faces a unique challenge: how can we actively select nodes for LLMs to annotate and consequently enhance the GNN training? How can we leverage LLMs to obtain annotations of high quality, representativeness, and diversity, thereby enhancing GNN performance with less cost? To tackle this challenge, we develop an annotation quality heuristic and leverage the confidence scores derived from LLMs to advanced node selection. Comprehensive experimental results validate the effectiveness of LLM-GNN. In particular, LLM-GNN can achieve an accuracy of 74.9% on a vast-scale dataset \products with a cost less than 1 dollar.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑