A Competition-based Learning Environment

CJ Chung

Associate Professor of Computer Science

Lawrence Technological University

Southfield, MI 48075

chung@LTU.edu

Tell me and I'll forget; show me and I may remember; involve me and I'll understand. -Chinese Proverb

I think one of the best ways to let students really involve in the learning subjects is to introduce classroom competitions. Classroom competitions are motivating and promoting students to work harder. Students will understand the classroom topic by doing. More importantly, they will learn real-world environments that are actually competitive, locally and globally.

Based on the above teaching philosophy, I have introduced some assignments as class room competitions in my classes. For example, in the fall 2006 semester in Intelligent Systems class, two of the 7 assignments were competitions.

The first competition was to design & implement an algorithm to guess intelligently n-digit secret numbers you have in mind. You, as a tester, are supposed to score just the number of correct digits in the correct position for each secret number. The learning objectives of this problem are to think about vast search space and to design a practically efficient search algorithm that can be run in polynomial time.

The winner of this first competition was Philip Munie (see a picture left). His program was flawless and passed all 4 test case files. The hardest case file contained 15 digit secret numbers. Average number of guesses for the 15 digit numbers was 49. He received a book as well as $10 bookstore gift card as prizes personally donated by the instructor.

The second place winner was Shawn Ellison. See a picture left. Even if his program failed in solving the 15 digit case, his algorithm outperformed that of Philip Munie. His average number of guesses for 12 digit numbers was 28 as compared to 31 by Philip. Shawn also received a book as a prize.

The second competition as the last assignment was to design, implement, and train an Artificial Neural Network (ANN) that will enable a robot to follow a solid or dashed line. Students were given web cameras to be used as a vision sensor for the laptop robot. As a specific requirement, students were asked to use ES (1+1) with 1/5 rule algorithm to train the ANN. Due to the limited class time in just one semester, students were not actually driving the robot, but displaying or animating the robot’s direction to go when it sees the line on the floor via the web camera. This assignment integrates many concepts learned in class such as Evolutionary Computation, ANN, Adaptability, and Computer Vision.

The winner prize (Best Buy $15 gift certificate personally donated by Dr. Chung) went to Gary Givental,  who passed all the different test cases perfectly. His training time was less than 5 seconds and he achieved near zero (0.1E-5) training error optimization. After the competition, he had a chance to share his ideas, know-how, and learning experience with his classmates.

Socrates tells that Education is the kindling of a flame, not the filling of a vessel. I was so pleased to see that students were really motivated from classroom competitions. Now it is time to think about the following what Geoff Nicholson, Former VP at 3M told us. Educating the entrepreneur mindset to students in order to win and survive in these global competitions in the real-world becomes important.

Research turns money into knowledge; Innovative entrepreneur mindset turns knowledge into money