COSCUP x RubyConf TW 2021

NLP 陪聊 AI 機器人的設計與使用體驗改進
08-01, 13:30–14:00 (Asia/Taipei), TR209
Language: 漢語


Translate Title

Design and UX research on AI chatbot

Abstract

講者的畢業專題,是探討「陪聊、聊天型」聊天機器人的設計,與使用者體驗的關係。該專題使用的機器人,是使用 AI 模型來生成回應的對話式機器人,在聊天時有不俗的表現。

在這場議程,講者希望分享他的 AI 陪聊機器人設計,包含如何訓練 AI?如何撰寫實驗頁面?以及至今的研究成果,即不同聊天機器人設計帶來的使用體驗影響。目前的改進方向,包含自訂大頭貼來激發用者正向情感,以及透過多回應讓使用者較難收不到適合的回應。

講者期待能藉由分享自己的研究拋磚引玉,與各位交流使用者體驗的研究方法,激發出更多改進使用體驗設計上的點子。

Talk Length

20

您是否知悉並同意如採遠端形式分享,需提供預錄影片(您需同意大會才能接受您的稿件) – yes Difficulty

入門

講者所屬的公司或組織名稱

學校:國立臺灣師範大學 / 實習公司:FiO

講者所屬社群

SITCON (?)

Target audience
  1. 經常思考如何讓 Chatbot 變得更好用、更讓使用者喜歡的開發者

  2. 對使用者體驗研究有興趣的人

  3. 對學術研究如何與程式開發結合有興趣的學生

English Abstract

The speaker’s graduation project is to explore the relationship between the design of a "chat-oriented" chatbot and the user experience. The bot used in this topic is a conversational chatbot that uses an AI model to generate responses and has a good performance in chatting.

In this agenda, the speaker hopes to share his AI chat robot design, including how to train AI? How to write an experiment page? And the research results so far, that is, the user experience impact of different chatbot designs. The current improvement directions include customizing photo stickers to stimulate users' positive emotions and making it less possible for users to not receive appropriate responses through provide multiple responses.

The speaker looks forward to sharing his own research result and sharing with you the research methods of user experience, then inspiring more ideas for improving user experience design.

other info

slido url

https://app.sli.do/event/zbagnecr

hackmd url

https://hackmd.io/@coscup/rkgpLTvCO/%2F%40coscup%2FSJTo8avRO