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 2402.16568cs.CL
收藏

基于大语言模型的时序知识图谱两阶段生成式问答

Two-stage Generative Question Answering on Temporal Knowledge Graph Using Large Language Models

National University of Defense Technology(国防科技大学);Meituan(美团);Xiangjiang Laboratory(湘江实验室)

Yifu Gao, Linbo Qiao, Zhigang Kan, Zhihua Wen, Yongquan He, Dongsheng Li

AI总结:

本文提出基于大语言模型的两阶段生成式时序知识图谱问答框架GenTKGQA,通过子图检索与答案生成,结合虚拟知识指示器融合图神经网络与文本表示,在两个数据集上验证了优越性。

摘要:

时序知识图谱问答(TKGQA)是一项极具挑战性的任务,原因在于问题中隐藏着时序约束,且需要从动态结构化知识中寻找答案。尽管大语言模型(LLM)在结构化数据推理能力上取得了显著进展,但其在 TKGQA 任务中的应用仍是一个相对未知的领域。本文首次提出了一种新颖的生成式时序知识图谱问答框架 GenTKGQA,该框架通过两个阶段引导 LLM 回答时序问题:子图检索和答案生成。首先,我们利用 LLM 的内在知识挖掘问题中的时序约束和结构链接,无需额外训练,从而在时序和结构两个维度缩小子图搜索空间。接着,我们设计了虚拟知识指示器,以非浅层的方式融合子图的图神经网络信号与 LLM 的文本表示,这有助于开源 LLM 通过指令微调深入理解检索到的事实之间的时序顺序和结构依赖关系。在两个广泛使用的数据集上的实验结果证明了我们模型的优越性。

英文摘要:

Temporal knowledge graph question answering (TKGQA) poses a significant challenge task, due to the temporal constraints hidden in questions and the answers sought from dynamic structured knowledge. Although large language models (LLMs) have made considerable progress in their reasoning ability over structured data, their application to the TKGQA task is a relatively unexplored area. This paper first proposes a novel generative temporal knowledge graph question answering framework, GenTKGQA, which guides LLMs to answer temporal questions through two phases: Subgraph Retrieval and Answer Generation. First, we exploit LLM's intrinsic knowledge to mine temporal constraints and structural links in the questions without extra training, thus narrowing down the subgraph search space in both temporal and structural dimensions. Next, we design virtual knowledge indicators to fuse the graph neural network signals of the subgraph and the text representations of the LLM in a non-shallow way, which helps the open-source LLM deeply understand the temporal order and structural dependencies among the retrieved facts through instruction tuning. Experimental results on two widely used datasets demonstrate the superiority of our model.

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。

↑