Beyond Static Question Banks: Dynamic Knowledge Expansion via LLM-Automated Graph Construction and Adaptive Generation

TL;DR

This work tackles two core bottlenecks in personalized education: costly manual knowledge-graph construction and insufficient state-aware reasoning over learner knowledge. It introduces Generative GraphRAG, pairing Auto-HKG for automated hierarchical knowledge-graph construction with CG-RAG for mastery-driven graph reasoning and retrieval-augmented generation of exercises. Key contributions include an end-to-end Auto-HKG pipeline with schema-constrained graph extraction, a CG-RAG component that integrates Bayesian Knowledge Tracing with graph-aware retrieval and LLM-based generation, and deployment evidence in real-world settings showing enhanced personalized learning experiences. The approach promises scalable, coherent, and adaptable learning paths across subjects by aligning content with both structured knowledge and evolving learner mastery.

Abstract

Personalized education systems increasingly rely on structured knowledge representations to support adaptive learning and question generation. However, existing approaches face two fundamental limitations. First, constructing and maintaining knowledge graphs for educational content largely depends on manual curation, resulting in high cost and poor scalability. Second, most personalized education systems lack effective support for state-aware and systematic reasoning over learners' knowledge, and therefore rely on static question banks with limited adaptability. To address these challenges, this paper proposes a Generative GraphRAG framework for automated knowledge modeling and personalized exercise generation. It consists of two core modules. The first module, Automated Hierarchical Knowledge Graph Constructor (Auto-HKG), leverages LLMs to automatically construct hierarchical knowledge graphs that capture structured concepts and their semantic relations from educational resources. The second module, Cognitive GraphRAG (CG-RAG), performs graph-based reasoning over a learner mastery graph and combines it with retrieval-augmented generation to produce personalized exercises that adapt to individual learning states. The proposed framework has been deployed in real-world educational scenarios, where it receives favorable user feedback, suggesting its potential to support practical personalized education systems.

Figures (4)

• Figure 1: Overview of the proposed Generative GraphRAG framework. The framework addresses two key challenges in personalized education systems: the high cost of manual knowledge graph construction, and the lack of learner-state-aware reasoning in tag-matching-based systems. Auto-HKG automatically constructs a hierarchical knowledge graph from question banks and textbooks, while CG-RAG integrates learner mastery states with graph-based reasoning and retrieval-augmented generation to produce personalized exercises.
• Figure 2: System architecture of the proposed framework, integrating automated hierarchical knowledge graph construction (Auto-HKG) with learner-aware graph reasoning and retrieval-augmented generation (CG-RAG) for personalized exercise recommendation.
• Figure 3: Visualization of the hierarchical educational knowledge graph automatically constructed by Auto-HKG.
• Figure 4: Visualization of the hierarchical educational knowledge graph automatically constructed by Auto-HKG.