Intelligent User Interfaces
Task-Based User Interface Project
Mikey: An Intelligent Tutor for Teaching Scientific Inquiry Skills
Michael A. Sao Pedro
Mikey is an intelligent tutor designed for gaining practice with various parts of the scientific method. Students form hypotheses, create experiments to test hypotheses, and formulate conclusions based on the data. The motivation is to improve elementary and middle school students' scientific inquiry skills. This project also serves as a stepping stone in my ongoing research for improving science education using Intelligent Tutoring Systems. See the Assistment Project at WPI for more information. The objective of the exercise is to conduct an experiment to determine what factors will cause a ball to roll farther away from a ramp or for a longer period of time. Students can modify the ramp steepness, ramp surface, the type of ball, and the run length.
The system allows students to create a hypothesis or let Mikey automatically generate one. Students also create their own ramp setups and see the ball roll down the ramp. Mikey critiques the trials set up by the student by checking if they correctly followed the control of variables (CVS) strategy. This strategy states that to test if a single variable affects an outcome, only that variable's value should change between trials. Mikey gives feedback for trials that do not change the variable of interest and trials in which other variables have changed. Mikey will inform the student of her mistake and either give her a verbal hint or highlight the controls that need to be changed, depending on the number of times each mistake was made. Mikey also computes the correctness of the conclusions stated by the student using her hypothesis and generated data. If an error in conclusions is detected, Mikey states what went wrong. If at any time the student does not know what to do next, she can ask Mikey for help by typing "next".
ANSI/CEA-2018 Task Model:
Executable Jar File ("java -jar SmartInclinedPlane_fat.jar"). Requires JDK 1.5+.
Source Code (Eclipse project file)
For more information about this project and/or the CEA-2018 standard, contact Professor Charles Rich.