Existing dialogue data collection methods such as the Wizard of Oz method (WoZ) or real dialogue recording are costly, and they prevent launching a new dialogue system. In this study, we requested crowd workers in crowdsourcing to create dialogue scenarios according to the instruction of the situation for persuasive dialogue systems that use emotional expressions. We collected 200 dialogues in 5 scenarios for a total of 1,000 via crowdsourcing. We also annotated emotional states and users' acceptance for system persuasion by using crowdsourcing. We constructed a persuasive dialogue system with the collected data and evaluated the system by interacting with crowd works. From the experiment, it was investigated that the collected labels have sufficient agreement even if we did not impose any training of annotation to workers.
@InProceedings{YOSHINO18.462, author = {Koichiro Yoshino and Yoko Ishikawa and Masahiro Mizukami and Yu Suzuki and Sakriani Sakti and Satoshi Nakamura}, title = "{Dialogue Scenario Collection of Persuasive Dialogue with Emotional Expressions via Crowdsourcing}", booktitle = {Proceedings of the Eleventh International Conference on Language Resources and Evaluation (LREC 2018)}, year = {2018}, month = {May 7-12, 2018}, address = {Miyazaki, Japan}, editor = {Nicoletta Calzolari (Conference chair) and Khalid Choukri and Christopher Cieri and Thierry Declerck and Sara Goggi and Koiti Hasida and Hitoshi Isahara and Bente Maegaard and Joseph Mariani and Hélène Mazo and Asuncion Moreno and Jan Odijk and Stelios Piperidis and Takenobu Tokunaga}, publisher = {European Language Resources Association (ELRA)}, isbn = {979-10-95546-00-9}, language = {english} }
Summary of the paper