Sequential and Session-Based Recommendation

CP4285 Modern Recommendation Systems — Week 05 QR code for the Week 05 slide deckScan for the slides
soc-n.us/cp4285-t2610-w05

CP4285 Instruction Team

08 Sep 2026

01 Overview

Week 5: Sequential and Session-Based Recommendation

Learning outcomes

  • Model temporal preferences.
  • Implement next-item prediction.
  • Compare static and dynamic representations.
  • Critically assess engagement-driven objectives.

02 Sequence Models

From sessions to next-item prediction

TODO: add GRU4Rec, SASRec, and Transformer recommender material.

Ethics Thread: Engagement objectives

Warning

TODO: examine manipulation risks in engagement optimization.

Summary

🔑 Time changes what a preference signal means.

Next week: retrieval and ranking architectures.