Teaching K-12 Math Online: Reasoning Mind CEO Alex Khachatryan (Part 4)

SM: Is your system entirely based on AI, or do you have teachers integrated into the process?

AK: In additional to the artificial intelligence system that would teach students, we have given teachers a very powerful control panel. They can get access to all kinds of just-in-time reports to see how each student is doing in the system. Teachers have access to diagnostic tools that tell them exactly what kinds of difficulties their students are having. Teachers can interfere on an as-needed basis and help students overcome difficulties in the moments when they are actually occurring. After the teacher intervention, the artificial intelligence system resumes course delivery.

SM: When did you start developing the system?

AK: The first ideas occurred in 2000. The first prototype was built in 2002. We spent a year developing a limited prototype and it only covered a couple of objectives in the middle school curriculum. It did not have all the functionality that we wanted it to have. However, the objective was to prove the concept, which it did. It worked phenomenally well.

SM: When you say it worked phenomenally well, what does that mean? How did you conduct your pilot?

AK: We went to two schools, one in Houston and the other in College Station. One had 30 students in the program and the second school had 20. We also set up control groups and did randomized control trials in both schools. We had an independent evaluator from the University of Houston who did the statistical analysis to evaluate the Reasoning Mind curriculum versus the regular curriculum. The report was extremely positive for Reasoning Mind.

There was statistical significance and educational significance differences on the order of 3 standard deviations, which is unheard of in an education environment. Anything that is more than one-third of a standard deviation is considered to be educationally significant, and we saw a standard deviation in the 2-3 range for educational performance for Reasoning Mind students versus standard curriculum students.

The feedback we received from kids and teachers was good as well. Kids particularly loved the system and were engaged. They really believed the system helped them tremendously in their effort to learn mathematics.

SM: In 2002, after you completed the proof of concept, what was the next step?

AK: The philanthropic community saw our results in a very positive way. We managed to raise more money. My wife and I had been working as volunteers for three years. We maintained our day jobs and we were doing this on weekends and at night. We hired three software engineers in Moscow to build the prototype, which is why it was done so cheap. I had to go to Moscow frequently to get this done.

By 2003 I realized that it was consuming too much of my time. I quit my day job and went 100% with Reasoning Mind, as did my wife. We were able to do that once we experienced some success and once we were able to raise some money. We also hired some people in Houston which allowed us to completely redesign the system and move away from the prototype into a production-grade system.

We developed a full grade curriculum that began at the end of elementary school. We wanted to get in touch with students early, before they developed significant gaps in their education.

This segment is part 4 in the series : Teaching K-12 Math Online: Reasoning Mind CEO Alex Khachatryan
1 2 3 4 5 6 7