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 2103.07102cs.CL
收藏

基于BERT的归纳关系预测

Inductive Relation Prediction by BERT

University of California, Santa Barbara(加利福尼亚大学圣巴巴拉分校)

Hanwen Zha, Zhiyu Chen, Xifeng Yan

摘要:

知识图谱中的关系预测主要由基于嵌入的方法主导,这些方法主要集中在转换设置上。不幸的是,他们无法处理归纳学习,其中存在看不见的实体和关系,并且无法利用先验知识。此外,他们的推理过程不容易解释。在这项工作中,我们提出了一个一体化的解决方案,称为BERTRL(BERT-based Relational Learning),它利用预先训练好的语言模型,并通过将关系实例及其可能的推理路径作为训练样本对其进行微调。在18种情况下,BERTRL在感应和传导设置中的15种情况下优于SOTA。同时,它在少样本学习中表现出很强的泛化能力,并且是可解释的。

英文摘要:

Relation prediction in knowledge graphs is dominated by embedding based methods which mainly focus on the transductive setting. Unfortunately, they are not able to handle inductive learning where unseen entities and relations are present and cannot take advantage of prior knowledge. Furthermore, their inference process is not easily explainable. In this work, we propose an all-in-one solution, called BERTRL (BERT-based Relational Learning), which leverages pre-trained language model and fine-tunes it by taking relation instances and their possible reasoning paths as training samples. BERTRL outperforms the SOTAs in 15 out of 18 cases in both inductive and transductive settings. Meanwhile, it demonstrates strong generalization capability in few-shot learning and is explainable.

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。

↑