Research Trends On Deep Learning Approaches In Education: A Bibliometric Analysis

Abstract

The deep learning approach in education, understood as meaningful, mindful, and joyful learning that cultivates higher-order thinking skills (HOTS), has increasingly attracted scholarly and policy attention. Unlike artificial intelligence-based deep learning, this pedagogical approach emphasizes student-centered learning, critical reasoning, creativity, and the integration of knowledge into real-life contexts. This study systematically maps research trends on deep learning approaches in education using bibliometric analysis. A dataset of 1,682 open-access documents (articles and conference papers) published between 2020 and 2025 was analyzed to identify patterns of publication growth, leading sources and affiliations, authorship collaboration, as well as thematic hotspots and emerging directions. The findings reveal a consistent upward trajectory in publication outputs with an annual growth rate of 22.98%, dominated by contributions from Asian institutions, particularly China, alongside notable global collaborations. Thematic analysis highlights strong intersections between deep learning approaches and areas such as engineering education, digital pedagogy, and curriculum innovation, while emerging topics point to adaptive learning, contrastive learning, and cognitive processes like short-term memory. These results suggest that the field is shifting from foundational exploration toward more integrative and applied pedagogical models. This bibliometric study provides new insights for educators, researchers, and policymakers to anticipate future developments and strengthen the role of deep learning approaches in fostering transformative and equitable educational practices.