Neural Recommendation Models

CP4285 Modern Recommendation Systems — Week 04 QR code for the Week 04 slide deckScan for the slides
soc-n.us/cp4285-t2610-w04

CP4285 Instruction Team

01 Sep 2026

01 Overview

Week 4: Neural Recommendation Models

Learning outcomes

  • Build neural recommenders.
  • Compare neural and latent-factor approaches.
  • Analyze explainability challenges.
  • Assess trade-offs between complexity and transparency.

02 Neural Models

Neural collaborative filtering

TODO: introduce model architectures, training, and comparisons.

Ethics Thread: Explainability and performance

Warning

TODO: examine transparency costs of increasingly complex models.

Summary

🔑 Neural capacity changes both performance and interpretability.

Next week: sequential and session-based recommendation.