Skip to main navigation Skip to search Skip to main content

Frequently Asked Question Pair Generation for Rule and Regulation Document

  • Keyang Ding
  • , Chenran Cai
  • , Shijue Huang
  • , Rui Wang
  • , Qianlong Wang
  • , Jianxin Li
  • , Guozhong Shi
  • , Feiran Hu
  • , Fengxin Li
  • , Ruifeng Xu*
  • *Corresponding author for this work
  • Harbin Institute of Technology
  • Harbin Institute of Technology Shenzhen
  • China Merchants Securities Co., Ltd.

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

Abstract

This paper presents a novel task to generate frequently asked question (FAQ) pairs for the rule and regulation documents. It offers an easy way for customers and employers to quickly gain knowledge of them and provides a potential corpus for question-answering robots. While previous work focuses on web texts (e.g., Wiki), we generate FAQ pairs from the formal and verbose rule and regulation documents, which is significant in real scenarios. To tackle this task, firstly, we carefully design a rules-based method to generate FAQ pairs based on structure information. Then we propose a pipeline framework for FAQ pair generation by deep learning. For experiments, we collect and annotate a Chinese FAQ pair generation dataset from documents of China Merchants Securities Co., Ltd. The results show that our method can generate proper FAQ pairs and achieve competitive performance in both automatic and human evaluation.

Original languageEnglish
Title of host publicationArtificial Intelligence and Mobile Services – AIMS 2022 - 11th International Conference, Held as Part of the Services Conference Federation, SCF 2022, Proceedings
EditorsXiuqin Pan, Ting Jin, Liang-Jie Zhang
PublisherSpringer Science and Business Media Deutschland GmbH
Pages49-63
Number of pages15
ISBN (Print)9783031235030
DOIs
StatePublished - 2022
Externally publishedYes
Event11th International Conference on Artificial Intelligence and Mobile Services, AIMS 2022 held as Part of the Services Conference Federation, SCF 2022 - Honolulu, United States
Duration: 10 Dec 202214 Dec 2022

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume13729 LNCS
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Conference

Conference11th International Conference on Artificial Intelligence and Mobile Services, AIMS 2022 held as Part of the Services Conference Federation, SCF 2022
Country/TerritoryUnited States
CityHonolulu
Period10/12/2214/12/22

Keywords

  • Deep learning
  • Frequently asked question pair generation
  • Natural language processing

Fingerprint

Dive into the research topics of 'Frequently Asked Question Pair Generation for Rule and Regulation Document'. Together they form a unique fingerprint.

Cite this