Interdisciplinary Knowledge Graph-Driven STEM Teacher Assistant Agent: Model Construction and Evaluation
Abstract
Interdisciplinary learning has emerged as a key priority in China's ongoing education reform, yet STEM teachers often face major challenges including fragmented disciplinary knowledge, limited collaboration, and insufficient experience in interdisciplinary lesson design. This study introduces an AI-powered STEM teacher assistant agent, underpinned by a triadic synergy among frontline STEM teachers, an 86,542-node interdisciplinary knowledge graph, and the DeepSeek-R1-671B foundation model. Within the LangChain framework, the system orchestrates prompt-engineered tasks, applies Chain-of-Thought reasoning for stepwise inference, and employs graph-based retrieval-augmented generation (GraphRAG) to deliver explainable, curriculum-aligned support for interdisciplinary instructional design and knowledge integration. To evaluate system efficacy, we developed SciCurriculum-QA, a 226-item benchmark aligned with China's national curriculum. Using the RAGAS framework, the SciKG-RAG model demonstrated superior performance across four key metrics: Faithfulness (0.8064), Answer Relevancy (0.8785), Context Precision (0.9467), and Context Recall (0.8702). Controlled experiments further demonstrated clear advantages over vector-based baselines in multi-hop reasoning and alignment with China's national curriculum. A qualitative study involving 21 in-service teachers confirmed perceived gains in instructional confidence, cross-domain coherence, and classroom applicability. Overall, the study…
Read the paper
Related studies
- Knowledge Management Perspective of Generative Artificial Intelligence
- Navigating Generative Artificial Intelligence Promises and Perils for Knowledge and Creative Work
- Characterizing the evolving research on enterprise content management
- Large language models and multimodal foundation models for precision oncology
- From Generative AI to Extended Reality: Multidisciplinary Perspectives on the Challenges, Opportunities, and Future of Educational Computing
- Generative Artificial Intelligence: A New Engine for Advancing Environmental Science and Engineering
Go deeper in Evidence Hub
Evidence Hub adds a study data sheet (design, sample size, evidence level, key numbers), an AI summary and a chat grounded in the paper's text.