Deep Learning in Educational Scenario

The relationship between deep learning and human neural networks is speculative due to limited understanding of brain processes. Deep Learning employs artificial neural networks to predict patterns in data, resembling biological brain functions. This connection extends to education, where AI models simulate learning processes. The overlap involves the dynamics of human learning and the attempt to replicate them through automated structures. Deep Knowledge Tracing, presented at NeurIPS 2015, utilizes recurrent neural networks to predict student performance based on previous data. In education, this approach predicts personalized training needs efficiently, fostering a significant human–machine relationship. The ASSISTments project exemplifies collaboration between researchers and teachers, offering a platform for interactive learning. The success is attributed to continuous repetition in science-related modes and teacher involvement. This reflects two educational data mining approaches: organic interaction between computers and humans and a more automated AI-mediated synthesis. Balancing AI advantages with human-centric education is crucial, emphasizing the need for a control policy and literacy strategy for intelligent technologies. The challenge is to develop intensive AI education forms in the coming months. This contribution explores operational strategies for implementing AI education, addressing challenges such as diverse Deep Learning algorithms, non-open datasets, theoretical hesitations, and BigTech interests. A focus on AI in education (AIED) involves training in Computational Thinking. The Korean experiment demonstrates successful Deep Learning education for children. The challenge lies in balancing technological literacy with ontological perspectives. Doleck et al. emphasize explaining rather than just predicting AI processes. Holmes et al. propose AIED applications like collaborative learning, student forum monitoring, continuous assessment, AI learning companions, and AI teaching assistants. The democratization of AI in education requires expanding datasets, broadening data sources, integrating real contexts, clarifying algorithms, and enhancing AI competence in curricula. Educational Data Mining (EDM) plays a crucial role in predicting student achievement. The article suggests a conceptual framework for AI integration, encompassing cognitive, biometric, physical, and spatial dimensions, alongside algorithmic, educational dataset, and subjective feedback considerations. Open-ended conclusions emphasize the need for a comprehensive, curriculum-based, and critical approach to AI in education, focusing on digital literacy, dataset understanding, and the impact of AI on subjective experiences.