Grounding proposition stores for question answering over linked data

作者：

Highlights：

• Grounding natural language is required for question answering but the contribution remains unmeasured.

• We provide a standalone evaluation and propose a method to ground propositions into a knowledge base.

• Results show how grounding accounts for 78.6.

• Simple lexical expansion can improve the results from 0.8.

摘要

•Grounding natural language is required for question answering but the contribution remains unmeasured.•We provide a standalone evaluation and propose a method to ground propositions into a knowledge base.•Results show how grounding accounts for 78.6.•Simple lexical expansion can improve the results from 0.8.

论文关键词：Question answering,Semantic parsing,Linked data,Grounding

论文评审过程：Received 17 June 2016, Revised 25 April 2017, Accepted 25 April 2017, Available online 26 April 2017, Version of Record 25 May 2017.

论文官网地址：https://doi.org/10.1016/j.knosys.2017.04.016