Tuesday, March 15, 2011

HLL: Artificial Intelligence for Math Tutorial

Visit the High Level Logic (HLL) Website
— Home of XPL (eXtensible Process Language)

It's like XML, but you can store source code in it.

I've emphasized the history of thought behind HLL. Recently a retired teacher sent me a copy of a letter that I sent to her more than 20 years ago. At the time, she was working with a software company and a programmer to create a mathematics tutorial game. She had asked for my input after I mentioned A.I. as a way to improve the teaching performance of tutorial games. I provided the following: August 27, 1989. Keep the HLL General Problem Solver in mind. Not only could it provide a detailed report to the human tutor, but that same mechanism could be used to automatically alter the tutorial.


The math tutor provides one on one instruction to the math student. The unique approach taken in the design of the software system is based on artificial intelligence techniques. These techniques are applied to continuously analyze the student's performance. Analysis is also done at a more detailed level than might usually be expected of a software system.

Continuous, detailed tracking of performance is required in order for the system to “decide” what specific skills, numbers, or operations the student is having trouble with, and how to structure tutorials or exercises to correct the difficulty. The student does not have to work through an entire set of problems, receive feedback in the form of a score, and then move up or down to a different level. Progress is encouraged by a system that continuously adapts to the student's needs, creating and presenting appropriate responses – customized to the individual needs – automatically, as the student works.

The approach helps to student progress in understanding mathematical concepts and become more competent in the application of mathematics to everyday problems. Boredom is abated by continuously keeping up with the student's abilities. Frustration is reduced by responding to problems as quickly as possible. In a sense, the use of artificial intelligence techniques allows the software to act a little more human by being more flexible, and responding more quickly to the student's needs. This explains the description of the software as a “one on one instructor.”

No comments:

Post a Comment