Presentation is loading. Please wait.

Presentation is loading. Please wait.

Personalizing conversational agent based e-learning applications

Similar presentations


Presentation on theme: "Personalizing conversational agent based e-learning applications"— Presentation transcript:

1 Personalizing conversational agent based e-learning applications
11/22/2018 2:49 AM Personalizing conversational agent based e-learning applications How can information about the user, and their affect, personality, performance, learning styles be used to improve the experience for a student interacting with a conversational agent in a learning activity? Mike Procter AU Graduate Student Conference 2015 © 2007 Microsoft Corporation. All rights reserved. Microsoft, Windows, Windows Vista and other product names are or may be registered trademarks and/or trademarks in the U.S. and/or other countries. The information herein is for informational purposes only and represents the current view of Microsoft Corporation as of the date of this presentation. Because Microsoft must respond to changing market conditions, it should not be interpreted to be a commitment on the part of Microsoft, and Microsoft cannot guarantee the accuracy of any information provided after the date of this presentation. MICROSOFT MAKES NO WARRANTIES, EXPRESS, IMPLIED OR STATUTORY, AS TO THE INFORMATION IN THIS PRESENTATION.

2 Conversational agents and students
CAs allow natural language interaction with regard to some domain of knowledge Animated pedagogical agents, Intelligent tutor systems, Game-base learning Role-playing actors Appearance of understanding, reply reasonably, follow rules of conversation, appear intelligent Non-verbal communication, emotional awareness, understanding personality, goals help maintain the illusion and can be used to direct “conversation”

3 Proposed solution

4 Student model data sources
DSA DSA DSA Student Model Agent DSA

5 Student model The DSA agents will form the essence of the model (Model = DSA agents + integration agent) The selection of data sources is based on what is available, and what the CA can use A base set will always be available. These will use the conversational record (because it’s always available). Engagement classifier based on ML Computational model of emotion

6 Evaluation An existing CA (Freudbot) will be tested
Students will chat with standard version of Freudbot, or one enhanced by the described system Chats will be followed up by a questionnaire to measure students impression of the experience Conversation logs will be analyzed for evidence of the effectiveness of intervention actions


Download ppt "Personalizing conversational agent based e-learning applications"

Similar presentations


Ads by Google