Research Area:  Machine Learning
This work presents a pilot study for the application of argument mining techniques in the context of argumentative Dialogue Systems. We extract the argument structure of an online debate and show how it can be utilized to generate artificial persuasive dialogues in an agent-agent scenario. The interaction between the agents is formalized as argument game and the resulting artificial dialogues are evaluated in a user study by comparing them to human generated ones. The outcomes indicate that the artificial dialogues are logically consistent and thus show that the use of the employed argument annotation scheme was successful.
Keywords:  
Argument Mining
Dialogue Systems
Machine Learning
Deep Learning
Author(s) Name:  Niklas Rach, Saskia Langhammer, Wolfgang Minker & Stefan Ultes
Journal name:  
Conferrence name:  9th International Workshop on Spoken Dialogue System Technology
Publisher name:  Springer
DOI:  10.1007/978-981-13-9443-0_12
Volume Information:  
Paper Link:   https://link.springer.com/chapter/10.1007/978-981-13-9443-0_12