GraphRAG textbook knowledge retrieval for grounded tutoring

Design · All designs
Evidence · no claims cited

Description

The system implements Retrieval-Augmented Generation via the GraphRAG framework, in which textbook material is represented as a knowledge graph to provide contextually relevant information to the Tutor Agent and course-generation pipeline. The authors chose GraphRAG over normal vector RAG on the argument that its graph structure better represents educational content relations for contextual tutoring. The Tutor Agent retrieves textbook data to ground its responses, and the course pipeline accesses GraphRAG data when constructing structured learning plans.

Design Implications

Context

Requirements

• Textbook material represented as a knowledge graph to provide structured educational content

Constraints

• The article notes future work will explore the potential of different RAG architectures

Target Learners

• mathematics students using the tutoring platform

Learning Goals

• grounding tutor responses and generated courses in course textbook content

Claims

Related Designs

• Chain-of-Concept: topic-graph organisation of past assessments replacing similarity-based retrieval
• Retrieval-augmented generative AI learning platform for construction education
• Intelligent digital textbook framework with LLM-powered structured dialogue for reading comprehension support
• RAG-based GenAI chatbot used as the AI-assisted learning support in the external validity study

Examples

-

Key Sources

• Jarosław A. Chudziak and Adam Kostka. (2025). AI-Powered Math Tutoring: Platform for Personalized and Adaptive Education. arXiv preprint. https://arxiv.org/abs/2507.12484

Page metadata

Field	Value
type	design
id	graphrag-textbook-retrieval-math-tutor
title	GraphRAG textbook knowledge retrieval for grounded tutoring
description	The system implements Retrieval-Augmented Generation via the GraphRAG framework, in which textbook material is represented as a knowledge graph to provide contextually relevant information to the Tutor Agent and cours...
status	review
generated	by process:settle-candidates on 2026-10-10
trust	unverified — no one has confirmed this page's content
sources	1 — jarosław-a-chudziak-and-adam-kostka-2025

Frontmatter is the page's machine-readable record — see the Schema & Guide.