arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2402.08170cs.LGcs.AI
收藏

LLaGA:大语言与图助手

LLaGA: Large Language and Graph Assistant

The University of Texas at Austin(德克萨斯大学奥斯汀分校);Snap Inc(Snap公司)

Runjin Chen, Tong Zhao, Ajay Jaiswal, Neil Shah, Zhangyang Wang

AI总结:

针对大语言模型难以处理图结构数据的问题,研究提出LLaGA模型,通过结构感知序列转换和投影器适配大语言模型输入,在多数据集多任务的监督及零样本场景下超越现有最优图模型。

摘要:

图神经网络(GNNs)推动了图结构数据分析的发展。近年来,GPT-4等大语言模型(LLMs)的兴起开启了深度学习的新纪元。然而,由于图结构难以转化为语言形式,将大语言模型应用于图数据面临独特挑战。为此,我们提出大语言与图助手(LLaGA),这是一种创新模型,可有效整合大语言模型能力以处理图结构数据的复杂性。LLaGA保留了大语言模型的通用特性,同时将图数据调整为与大语言模型输入兼容的格式。其实现方式是将图节点重组为具备结构感知的序列,再通过多功能投影器将这些序列映射到词嵌入空间。LLaGA在通用性、泛化性和可解释性方面表现出色,能在不同数据集和任务上保持稳定性能,可将能力扩展到未见过的数据集或任务,还能为图提供解释。我们在主流图基准上开展的大量实验表明,单个LLaGA模型在四个数据集、三项任务上均表现优异,在监督和零样本场景下均超越了当前最优的图模型。我们的代码可在https://github.com/VITA-Group/LLaGA获取。

英文摘要:

Graph Neural Networks (GNNs) have empowered the advance in graph-structured data analysis. Recently, the rise of Large Language Models (LLMs) like GPT-4 has heralded a new era in deep learning. However, their application to graph data poses distinct challenges due to the inherent difficulty of translating graph structures to language. To this end, we introduce the Large Language and Graph Assistant (LLaGA), an innovative model that effectively integrates LLM capabilities to handle the complexities of graph-structured data. LLaGA retains the general-purpose nature of LLMs while adapting graph data into a format compatible with LLM input. LLaGA achieves this by reorganizing graph nodes to structure-aware sequences and then mapping these into the token embedding space through a versatile projector. LLaGA excels in versatility, generalizability and interpretability, allowing it to perform consistently well across different datasets and tasks, extend its ability to unseen datasets or tasks, and provide explanations for graphs. Our extensive experiments across popular graph benchmarks show that LLaGA delivers outstanding performance across four datasets and three tasks using one single model, surpassing state-of-the-art graph models in both supervised and zero-shot scenarios. Our code is available at \url{https://github.com/VITA-Group/LLaGA}.

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。

↑