.png){kind=icon}
返回
Knowledge base question answering via path matching
.png){kind=icon}
PRE
AI
DOI：10.1016/j.knosys.2022.109857.png)
.png){kind=icon}
原文求助.png){kind=icon}
分享.png){kind=icon}
收藏
摘要
En 中文
Knowledge base question answering (KBQA) refers to combining the information in the knowledge base to obtain an answer for an objective question. However, most of the existing methods tend to add a large number of hand-crafted features or constraints to improve the performance of the model. This makes the KBQA system more and more complex, but the improvement is limited. In this work, we try a novel method without any hand-crafted features or constraints, which transforms this problem into a text matching problem. In this method, a question and a series of edges in knowledge bases (KBs) are treated as two pieces of text that need to be matched. The network to match the two texts called Path Matching Model (PMM). On the WebQuestions benchmark, our method has 3% improvement compared to the state-of-the-art method.(c) 2022 Published by Elsevier B.V.
Keyword:
KBQA
Text matching
Knowledge graph
期刊
K
IF:
7.6
论文数:
1.3W
被引数:
4.5W
机构
引用论文
Intense pulsed light improves signs and symptoms of dry eye disease due to meibomian gland dysfunction: A randomized controlled study.png)
PLOS ONE
OAAI
PLOS ONEIF0
OAAI分享收藏
查看更多内容