top of page

Projects

Research Projects
AI-powered grading system - AutoGrade
Project Investigator

Background
This project explores the integration of Large Language Model (LLM), such as ChatGPT, as a powerful tool for grading assignments within learning management platforms like Canvas. Traditional grading methods often lack the precision and specificity that students require for a deeper understanding of their work. In contrast, ChatGPT, when equipped with appropriate prompts and programmed by instructors, provides a solution that can enhance the grading and feedback process. 

This approach enabled the provision of comprehensive feedback aligned with each item of the predefined rubric, facilitating students in pinpointing their strengths and areas for enhancement with precision, ultimately enhancing the overall learning experience. The AI tool can significantly reduce instructors' grading workload by automating assignment evaluation and offering detailed feedback based on predefined criteria.


Goal

Our team has developed an online platform incorporating GPT 4.0 to aid instructors in grading students' assignments efficiently. Throughout this endeavor, we aim to investigate the attitudes, expectations, and concerns of both students and instructors regarding the implementation of this AI grading platform.

The screenshot of the grading application

grading1.png

Feedback provided by AI with encouraging words.

image.png

Feedback provided by AI with detailed feedback that matched the specific criteria we had set

image.png

Research Projects
A New Method Using LLMs for Keypoints Generation in Qualitative Data Analysis

 

The screenshot of analysis results:

image.png

03 Research Projects
Work-on progress: Measuring Prosocial Behavior in Schools through a Virtual Reality Game: vSchool
 UX researcher/Game Designer

Background
The project of Vschool is to develop and validate a school-friendly, game-based assessment of prosocial behavior for 4th–6th-grade students. The assessment will involve a virtual interactive environment.
Goal

The goal of this project is to ensure the game interface is functional, engaging, and aligned with the assessment goals for target users.

Methods
The following methods were used in this project: usability test, a cognitive walkthrough of the game, System Usability Scale (SUS), eye tracking, and skin conductance response, open-ended survey, interview. 

Teacher Interview

 Student Interview

Classroom Observation
Results

Still work on this…

​

​

​

​

If you are interested in the game tests, please contact fyu@mail.missouri.edu

bottom of page