New Potentials for Data‐Driven Intelligent Tutoring System Development and Optimization

Abstract

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.

Adaptive structure metrics for automated feedback provision in intelligent tutoring systems

“…Most of the time, students proceed through a curriculum of different tasks to obtain skills and knowledge. The systems job is to select the next task depending on the current level of knowledge and individual parameters of the student (outer loop) and to support her solving the current task (inner loop) [5]. Such systems have been successfully applied in many contexts, especially in learning logic and math concepts, and have been proven to lead to positive learning outcomes for students [12,6].…”

“…Such systems have been successfully applied in many contexts, especially in learning logic and math concepts, and have been proven to lead to positive learning outcomes for students [12,6]. However, they usually rely on extensive knowledge engineering to formalize domain concepts and explicitly track student knowledge, which is both costly and difficult, especially in domains where explicit and detailed knowledge about the domain can not be obtained (so-called ill-defined domains) [3,4,5]. To relieve ITS engineers from the burden of knowledge engineering, data-driven approaches have emerged.…”

“…However, researchers have reported 100 -1,000 hours of authoring time for one hour of instructions in ITSs [3]; in addition, ITSs usually require an underlying domain theory such that their applicability is limited in areas where problems and their solution strategies are not easy to formalize [4,5]. In such domains, data-driven approaches are possible, providing feedback based on a set of existing examples for (correct) solutions of the underlying task [5,6]: If the students requires a hint on how to change her attempt to get closer to a correct solution, it can be compared to a similar example from the set, and the dissimilarities between her attempt and the example can be contrasted or highlighted in order to help the student to improve her own solution [7,8,9].…”

An Extensible Domain-Specific Language for Describing Problem-Solving Procedures

“…An approach that is recently gaining more and more attention is to use historical student data for developing an ITS [14]. Successful solutions from the past can be used to provide feedback and hints for students in the present, which circumvents the need to create an expert model.…”

“…In particular, there are model-tracing tutors [2], example-tracing tutors [1], constraint-based tutors [16], and data-driven tutors [14]. A significant di↵erence between the paradigms lies in the way the expert knowledge necessary for following the steps of a student, is specified.…”