Natural language understanding for argumentative dialogue systems in the opinion building domain

作者：

Highlights：

• We propose natural language understanding framework for argumentative dialogue systems

• We propose effective way of recognizing the user intent and identifying system arguments

• We have employed pre-trained BERT model and BiLSTM for intent and argument detection

• We collect user utterances for argumentation system in an extensive online study

• The evaluation indicates advantage of the utilized techniques over baseline models

摘要

•We propose natural language understanding framework for argumentative dialogue systems•We propose effective way of recognizing the user intent and identifying system arguments•We have employed pre-trained BERT model and BiLSTM for intent and argument detection•We collect user utterances for argumentation system in an extensive online study•The evaluation indicates advantage of the utilized techniques over baseline models

论文关键词：Natural language understanding,Intent classification,Sentence similarity,Argumentative dialogue system,Human–computer interaction

论文评审过程：Received 18 January 2021, Revised 27 November 2021, Accepted 25 January 2022, Available online 7 February 2022, Version of Record 16 February 2022.

论文官网地址：https://doi.org/10.1016/j.knosys.2022.108318