RAGEAR: Retrieval-Augmented Graph-Enhanced Academic Recommender

2026-05-26

AI Summary

This work addresses limitations in existing course recommendation systems, which predominantly rely on metadata and struggle to leverage fine-grained instructional content from lecture transcripts while inadequately modeling structured academic constraints such as credit requirements and prerequisites. To overcome these challenges, the authors propose a neuro-symbolic architecture that integrates dense retrieval with a knowledge graph: it retrieves semantically relevant lecture segments using full transcript context and encodes relationships among courses, lectures, credits, and curriculum plans via a graph structure. A graph-aware aggregation function propagates segment-level evidence to the course level, jointly considering the proportion, ranking strength, and distribution of relevant segments. Experiments on 152 student queries demonstrate that the proposed method significantly outperforms baselines using only metadata or lecture transcripts, substantially improving top recommendation quality as measured by the RAGEAR metric.

Abstract

We present RAGEAR (Retrieval-Augmented Graph-Enhanced Academic Recommender), a neurosymbolic recommender system for academic course recommendation. RAGEAR combines dense retrieval over full lecture transcripts with a symbolic Knowledge Graph modelling courses, lessons, transcript chunks, credits, study plans, and curricular information. The Knowledge Graph supports symbolic filtering and contextualisation based on structured constraints, such as credits, academic disciplines, study plans, and prerequisites. Unlike metadata-based approaches, it exploits fine-grained instructional content by retrieving transcript chunks semantically aligned with a student's query. The main contribution is a graph-aware aggregation function that propagates chunk-level evidence to course-level recommendations. The score combines three factors: the share of retrieved similarity associated with a course, the rank-based strength of its relevant chunks, and the distribution of evidence across lessons. We evaluate RAGEAR on 152 student-like queries through a human evaluation sample and a large-scale LLM-based relevance assessment. Results show that lecture transcripts improve over metadata-only retrieval, and that RAGEAR further improves ranking quality over a transcript-based normalized SumP baseline, especially for top-ranked recommendations.

Problem

• academic course recommendation
• lecture transcripts
• knowledge graph
• retrieval-augmented recommendation
• structured constraints

Innovation

• Retrieval-Augmented
• Graph-Enhanced
• Neurosymbolic Recommender
• Knowledge Graph
• Lecture Transcript Retrieval