Graph-Based Recommendation

CP4285 Modern Recommendation Systems — Week 09 QR code for the Week 09 slide deckScan for the slides
soc-n.us/cp4285-t2610-w09

CP4285 Instruction Team

13 Oct 2026

01 Overview

Week 9: Graph-Based Recommendation

Learning outcomes

  • Represent recommendations as graph problems.
  • Explain graph propagation.
  • Build graph-based recommenders.

  • Analyze risks of graph-driven feedback loops.

02 Graph Models

User-item graphs and LightGCN

TODO: add embeddings, propagation, and examples. ## Ethics Thread: Community amplification ::: {.callout-warning} TODO: examine homophily, echo chambers, and feedback loops. ::: —

Summary

🔑 Graph structure can reveal—and reinforce—connections.

Next week: multi-objective recommendation.