Deep Learning-Based Adaptive Learning System Innovation for Facilitating Diverse Curiosity in Upper Elementary Student

Abstract

The cultivation of curiosity in upper elementary education represents a critical yet underexplored frontier for deep learning-based adaptive systems. While deep learning has revolutionized knowledge tracing and personalized learning pathways, the systematic integration of curiosity as a multidimensional construct into adaptive system architectures remains largely absent from the scholarly literature. This systematic literature review examines the extant research landscape at the intersection of deep learning-based adaptive learning systems and curiosity facilitation for upper elementary students (Grades 4–6). Following PRISMA guidelines, a comprehensive search of Scopus, Web of Science, and IEEE Xplore databases yielded 1,847 initial records, from which 47 empirical studies met rigorous inclusion criteria. The findings reveal three critical research gaps: (1) the predominant focus on knowledge state modeling in deep knowledge tracing systems to the exclusion of epistemic curiosity constructs; (2) the absence of validated curiosity-aware student modeling architectures capable of distinguishing between and responding to diverse curiosity subtypes; and (3) the lack of pedagogical frameworks that operationalize curiosity-driven learning mechanisms within deep learning-based adaptive system designs. The review culminates in a proposed Curiosity-Informed Adaptive Deep Learning (CIADL) Framework, representing the first integrative model that systematically maps the theoretical, measurement, architectural, and pedagogical dimensions necessary for designing adaptive systems that foster, rather than merely accommodate, student curiosity. This framework establishes a novel research agenda for educational technology, with significant implications for the design of developmentally appropriate adaptive learning environments