Design and practice of high school STEM curriculum integrating few-shot learning and multi-agent systems

Abstract

In recent years, STEM education has received increasing attention in global education reform, emphasizing the integration of interdisciplinary knowledge and the ability to solve real-world problems. The development of problem-solving skills requires a combination of procedural and conceptual knowledge. Although example-based learning has been proven to be an effective instructional strategy, students often face challenges such as limited access to example resources and delayed feedback during practice. Multi-agent systems, with their collaborative, adaptive, and distributed characteristics, can provide continuous and personalized learning support for students. Based on this, the present study designed and developed a STEM curriculum integrating example-based learning and multi-agent systems, which has been successfully implemented in high school classrooms. This curriculum design offers practical guidance and a reference for promoting the deep integration of artificial intelligence and STEM education at the K–12 level.