"Increasing widespread use of educational technologies is producing vast amounts of data. Such data can be used to help advance our understanding of student learning and enable more intelligent, interactive, engaging, and effective education. In this article, we discuss the status and prospects of this new and powerful opportunity for data-driven development and optimization of educational technologies, focusing on intelligent tutoring systems We provide examples of use of a variety of techniques to develop or optimize the select, evaluate, suggest, and update functions of intelligent tutors, including probabilistic grammar learning, rule induction, Markov decision process, classification, and integrations of symbolic search and statistical inference."
Koedinger, Kenneth R., Emma Brunskill, Ryan Shaun Joazeiro de Baker, Elizabeth A. McLaughlin and John C. Stamper. “New Potentials for Data-Driven Intelligent Tutoring System Development and Optimization.” AI Magazine 34 (2013): 27-41.
1. SUMMARY
The paper discussed machine-learning (ML) and data-driven intelligent tutoring system development through the case of SimStudent and Hint Factory, to be followed by discussions on several optimization techniques.
SimStudent allows non-AI programmers to build the central cognitive component of an ITS by “tutoring” the model using tools like CTAT. The author enters tasks into the system, demonstrates and giving feedback. As a result, production rules for accomplishing the next steps (in an inner loop) can be induced. AI and ML learn this rule-based production system along the way.
HintFactory “automatically generating context-specific hints by using previously collected student data” (Barnes and Stamper 2008). This method targets students’ problem-solving capability, providing helps on demand and as specific as possible. It relies on a Markov-based action graph with rewards. Successor states with the highest reward values will be used to generate hint sequences
Optimization techniques include: an automated search process by “hypothesizing alternative knowledge representations and testing them against data” leveraging algorithms like the Learning Factor Analysis; the use of a statistical model to estimate and track student learning over the curriculum such as the knowledge tracing by Corbett and Anderson; and finally the use of AI and ML to measure difficult yet important factors like student engagement and affect.
2. STRENGTHS
The paper is a great, detailed survey on past research efforts in using Artificial Intelligence and Machine Learning to quickly build and optimize Intelligent Tutoring Systems.
3. WEAKNESSES
The only weakness is the lack of original work which is understandable since this is intended to be a survey/case study paper.
