Virtual Tutor - Department of Computer Science

Virtual Tutor Application v1.0
Ruth Agada
Dr. Jie Yan
Bowie State University
Computer Science Department
Background
• Animated agents have many uses.
▫ in educational, training, and simulation environments,
 HLS, virtual simulation for dismounted soldiers, etc
▫ and in entertainment applications.
 Second life, Sims, etc
• impression of realism.
Background cont’d
• Examples:
▫ Virtual Sign Animated Pedagogical Agents
 Designed for deaf and hearing impaired people.
 Presents sign demonstrations via a series of 3-D graphical
animations.
▫ Language learning Animated Pedagogical Agents
 Designed to teach children to read
 monitors each student’s performance continuously and
adapts the study plan
 Capable of providing feedback to students
Introduction
• Studies have shown that effective individual tutoring
is the most powerful mode of teaching.
▫ Logically and financially unfeasible.
• Animated pedagogical agents offer enormous
promise for interactive learning environments.
• Despite the progress, much remains to be done.
Introduction cont’d
• Research objective
▫ To design, develop and implement a virtual tutor
 To develop a powerful new approach for investigating
engaging and effective communication
 To develop an environment for supplemental learning
purposes.
 Create realistic 3D animated model to be used as virtual
tutor
Approach
• Interface design
▫ The interface consists of three main parts:
 an overview window.
 the agent window.
 the content explanation window.
▫ Quiz interface components:
 Question and answer area
 Agent window
 Report window
▫ Audio and phoneme files generated using tts
 Will be changed to a human voice
Approach cont’d
• Generating audio and phoneme files
▫ Use avatar-CUAnimate toolkit
 Using the text-to-speech component
▫ Tts file – 2 separate files
 Simple .wav file
 Phoneme .txt file
Avatar- CUAnimate toolkit
Lecture screenshot
Lecture screenshot with illustration
Quiz Screenshot
Other models
Are just a few of the
models
Evaluation
• Preliminary testing of Agent
▫ Tested on Bowie State University students.
▫ Observing subject’s perception.
▫ To assess student:
 12 survey questions on a 5 point Likert scale.
 Rated with a score range 1 – 5.
Sample Questionnaire
Subject Test Results Comparisons
5.5
Average Score
5.0
4.5
4.0
No Emotions, No
Head Movements
3.5
3.0
No Emotions, Head
Movements
2.5
2.0
1.5
1.0
Emotions, Head
Movements
Conclusion & future work
• Virtual Tutor can be used to supplement/teach lecture
materials.
▫ Students will then be able to view the prepared lectures
• Preliminary testing proved hypotheses.
• New features will be added to application
▫ More modules developed
▫ Watch instructional videos
▫ User can keep track of grades
• Make available online.
Acknowledgements
Thanks to:
Dr. Yan .
Department of Computer science.
Research made possible through grants from:
NSF: SGER: Research to Improve Communication
by Pedagogical Agents (Award No. IIS-0827188).