Personalized learning with dynamic knowledge graphs: a systematic approach in the context of smart cities

Personalized learning has become a critical research focus in modern education, aiming to enhance learning outcomes and engagement by tailoring educational content to individual learners. With the rise of smart cities, where data-driven technologies are transforming urban management and service delivery, personalized learning systems can also benefit from such technological advancements. This study proposes an innovative dynamic knowledge graph system to address the limitations of traditional learning frameworks in adaptability and personalized support within the context of smart cities. The system is designed with a three-layer architecture comprising the data layer, model layer, and application layer. The data layer employs multimodal data processing techniques to generate high-quality inputs, laying a robust foundation for system functionality. The model layer integrates knowledge graph construction, dynamic updates through graph neural networks (GNN), and reinforcement learning-driven learning path optimization, enabling dynamic knowledge management and decision-making. The application layer delivers precise personalized recommendations and adaptive learning paths through a closed-loop feedback mechanism. The technological innovations of this system are reflected in its real-time adaptability, high precision in resource recommendations, and dynamic learning path optimization capabilities. By deeply integrating graph neural networks and reinforcement learning, the system overcomes technical bottlenecks in flexibility and intelligence that challenge traditional personalized learning systems. As part of the smart city ecosystem, this system has extensive application potential across K-12 education, higher education, professional training, and both blended and online learning environments. However, challenges such as computational performance and scalability in large-scale applications remain. Future research will focus on improving the performance of core algorithms, exploring broader educational scenarios in smart cities, and enhancing user interaction to further strengthen the system’s value in personalized learning.