LIT-GRAPH: Evaluating Deep vs. Shallow Graph Embeddings for High-Quality Text Recommendation in Domain-Specific Knowledge Graphs

TL;DR

This paper compares deep and shallow graph embedding methods for recommending pedagogically relevant literature in domain-specific knowledge graphs, finding that deep models excel in semantic ranking while shallow models perform better in structural link prediction.

Contribution

It introduces LIT-GRAPH, a novel system for literature recommendation in education, and provides a comprehensive comparison of embedding paradigms highlighting their strengths.

Findings

Shallow models perform better in structural link prediction.

Deep R-GCN models excel in semantic relevance ranking.

Relation-specific message passing improves pedagogical relevance.

Abstract

This study presents LIT-GRAPH (Literature Graph for Recommendation and Pedagogical Heuristics), a novel knowledge graph-based recommendation system designed to scaffold high school English teachers in selecting diverse, pedagogically aligned instructional literature. The system is built upon an ontology for English literature, addressing the challenge of curriculum stagnation, where we compare four graph embedding paradigms: DeepWalk, Biased Random Walk (BRW), Hybrid (concatenated DeepWalk and BRW vectors), and the deep model Relational Graph Convolutional Network (R-GCN). Results reveal a critical divergence: while shallow models excelled in structural link prediction, R-GCN dominated semantic ranking. By leveraging relation-specific message passing, the deep model prioritizes pedagogical relevance over raw connectivity, resulting in superior, high-quality, domain-specific recommendations.