Improved Performances and Motivation in Intelligent Tutoring Systems: Combining Machine Learning and Learner Choice

Abstract: Large class sizes challenge personalized learning in schools, prompting the use of educational technologies such as intelligent tutoring systems. To address this, we present an AI-driven personalization system, called ZPDES, based on the Learning Progress Hypothesis - modeling curiosity-driven learning - and multi-armed bandit techniques. It sequences exercises that maximize learning progress for each student. While previous studies demonstrated its efficacy in enhancing learning compared to hand-made curricula, its impact on student motivation remained unexplored. Furthermore, ZPDES previously lacked features allowing student choice, a limitation in agency that conflicts with its foundation on models of curiosity-driven learning. This study investigates how integrating choice, as a gamification element unrelated to exercise difficulty, affects both learning outcomes and motivation. We conducted an extensive field study (265 7-8 years old children, RCT design), comparing ZPDES with and without choice against a hand-designed curriculum. Results show that ZPDES improves both learning performance and the learning experience. Moreover adding choice to ZPDES enhances intrinsic motivation and further strengthens its learning benefits. In contrast, incorporating choice into a fixed, linear curriculum negatively impacts learning outcomes. These findings highlight that the intrinsic motivation elicited by choice (gamification) is beneficial only when paired with an adaptive personalized learning system. This insight is critical as gamified features become increasingly prevalent in educational technologies.

• The paper introduces the ZPDES algorithm that uses multi-armed bandit techniques to personalize exercise selection for maximal learning progress.
• It evaluates the impact of integrating learner choice, showing that self-directed pathways significantly boost intrinsic motivation when combined with adaptive curricula.
• A field study with 265 students confirms that personalized, adaptive practice outperforms traditional, expert-designed curricula in enhancing learning outcomes.

Improved Performances and Motivation in Intelligent Tutoring Systems

The paper "Improved Performances and Motivation in Intelligent Tutoring Systems: Combining Machine Learning and Learner Choice" presents a comprehensive study on enhancing Intelligent Tutoring Systems (ITS) through the integration of machine learning techniques and the inclusion of student agency. The paper explores the implications of personalized learning paths in educational settings, particularly in addressing the challenges posed by high classroom sizes and the varying abilities of students.

The research leverages the Learning Progress hypothesis (LPH), focusing on maximizing learning progress through multi-armed bandit machine learning techniques. The core contribution of the study is the development of the ZPDES algorithm, which aims to personalize learning experiences by selecting exercises that offer maximal learning progress for students. While previous studies have documented the efficacy of ZPDES in improving learning performances across diverse student profiles compared to expert-designed curriculums, this paper addresses two limitations: the lack of assessment on motivational impact and the absence of student choice in learning paths.

To overcome these limitations, the authors introduced ZCO, an ITS that integrates adaptive exercise recommendations with the potential for students to express preferences, particularly in areas unrelated to exercise difficulty. This design choice aligns with the organization's overarching objective of linking intrinsic motivation with learning performance.

The paper's notable outcomes stem from an extensive field study involving 265 children from 11 schools. The findings endorse the hypothesis that personalization based on Learning Progress enhances learning outcomes and induces a positive and motivating learning experience. Moreover, the study demonstrates the significant role of self-choice in reinforcing intrinsic motivation, thereby amplifying the effectiveness of the personalized curriculum. Crucially, the presence of self-choice was shown to have a beneficial impact only when coupled with an efficient adaptive curriculum like that provided by LP-based learning.

Key observations suggest that while gamification elements can elicit intrinsic motivation, they do not inherently improve learning outcomes without an effective curriculum personalization strategy. Thus, the strategic deployment of gamified features should be carefully considered within the context of adaptive educational technologies.

The implications of this research are considerable for both theoretical and practical applications. Theoretically, it extends the LPH by validating the link between intrinsic motivation and learning progress in ITS contexts. Practically, it informs the development of future ITS that can dynamically adapt to learners' needs while simultaneously supporting their motivation and engagement through choice and personalized paths.

Future research could explore broader domains and age groups to enhance the generalizability of these findings. Additionally, the integration of cognitive and affective states in further ITS enhancements promises an intriguing direction for future studies. The paper underscores the significance of designing educational technologies that not only cater to varying learner profiles but also support sustained motivation, ultimately fostering better academic outcomes.