以深度學習技術發展交談機器人應用於LINE平台之初步探討

The Pilot Study for Deep Learning Technique Applied in Chatbot in LINE with Course of Negotiation Skills

黃展鵬(Jan-Pan Hwang) ； 張美瑤(Mei-Yao Chang)

《工程、技術與科技教育學術研討會》 第八屆 (2019/05) Pp. 241-249

https://doi.org/10.6571/SCETE.201905.0019

摘要

本研究結合深度學習的RNN／LSTM來發展一個交談式機器人，主要透過WebAPI結合Line Messaging API即時通訊的方式，應用於協商談判課程之中。在系統建置之後，於台灣某技職校院進行協商談判課程中的學習活動。主要應用於學生演練談判情境之對話，並針對系统使用於課程之可行性進行調查。在二週共四小時的課程學習活動之中，共有42位大學二、三年級且分佈不同學院和系別的學生。調查結果顯示，大多數學生對於使用LINE平台為做對話演練的方式抱有正面的看法。以此研究結果做為人工智慧技術應用於課堂教學之創新先例。

關鍵字

深度學習 ； 協商談判 ； 交談機器人 ； 即時通訊

並列摘要

This research combines the artificial intelligence technology RNN/LSTM to develop a dialogue system, which is mainly used in the bargaining unit of negotiation course. After the system was built, the learning activities were conducted in a vocational school in the south of Taiwan. It is mainly used in the dialogue of student drills in bargaining situations, and investigates the feasibility of the system used in the course. The two weeks course has 42 second and third-year university students with different colleges and departments. The results show that most students have a positive view of instant messaging conducts dialogue exercises, and feel interesting. The groups targeting dialogue robots, students prefer courses. For groups with real people, students think it is more helpful for learning, and they have higher satisfaction with the results of bargaining negotiations.

並列關鍵字

Deep learning ； Negotiation ； Chat-bot ； Instant messaging