@inproceedings{4495901ed417488eaafbf8a15324c033,
title = "Question answering over knowledge base with symmetric complementary attention",
abstract = "Knowledge Base Question Answering (KBQA), which aims to answer natural language questions with structured data from a knowledge base is an important Natural Language Processing (NLP) problem. To answer the question, we need to find the fact from the Knowledge Base whose subject and relation best match the question. Most existing methods treat this task as a pipeline of two separate subtasks: subject matching and relation matching. While ignoring the relevance between them. In this paper, we focus on solving this problem through a joint learning method. We present a neural joint model with a shared encoding layer to learn the two subtasks together to improve each other. In particular, we design a Symmetric Bidirectional Complementary Attention module based on the attention mechanism and the gate mechanism to model the relationship between the two subtasks. The experimental results demonstrate that our approach can obtain higher accuracy than the state-of-the-art method.",
keywords = "Joint learning, Knowledge base, Question answering",
author = "Yingjiao Wu and Xiaofeng He",
note = "Publisher Copyright: {\textcopyright} Springer Nature Switzerland AG 2020.; 7th International Workshop on Big Data Management and Service, BDMS 2020, 6th International Symposium on Semantic Computing and Personalization, SeCoP 2020, 5th Big Data Quality Management, BDQM 2020, 4th International Workshop on Graph Data Management and Analysis, GDMA 2020, 1st International Workshop on Artificial Intelligence for Data Engineering, AIDE 2020, held in conjunction with the 25th International Conference on Database Systems for Advanced Applications, DASFAA 2020 ; Conference date: 24-09-2020 Through 27-09-2020",
year = "2020",
doi = "10.1007/978-3-030-59413-8\_2",
language = "英语",
isbn = "9783030594121",
series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",
publisher = "Springer Science and Business Media Deutschland GmbH",
pages = "17--31",
editor = "Yunmook Nah and Chulyun Kim and Kim, \{Seon Ho\} and Yang-Sae Moon and Whang, \{Steven Euijong\}",
booktitle = "Database Systems for Advanced Applications. DASFAA 2020 International Workshops - BDMS, SeCoP, BDQM, GDMA, and AIDE, Proceedings",
address = "德国",
}