Home

Track

Recommender Systems

Goal: Learn ranking and retrieval for personalized recommendations, from classical collaborative filtering through matrix factorization, deep CTR models, two-stage industrial systems, and sequential recommenders.

Prereqs: ML Basics. Deep Learning helps for the neural ranking and retrieval papers.

Status: done

Work through the steps in order. Bold links open YouTube.

Step Concept YouTube Read
1 Recommendation problem and candidate stages The Math Behind Recommender Systems Google Developers — Recommendation systems overview
2 Content-based filtering Content-Based Recommendations Google Developers — Content-based filtering
3 Collaborative filtering Recommendation Systems — Collaborative Filtering Google Developers — Collaborative filtering
4 Item-based neighborhood methods   Sarwar et al. 2001 — Item-Based Collaborative Filtering Recommendation Algorithms
5 Matrix factorization How does Netflix recommend movies? Matrix Factorization Koren, Bell & Volinsky 2009 — Matrix Factorization Techniques for Recommender Systems
6 Implicit feedback and BPR BPR: Bayesian Personalized Ranking from Implicit Feedback Rendle et al. 2009 — BPR: Bayesian Personalized Ranking from Implicit Feedback
7 Wide & Deep memorization and generalization Wide & Deep Learning for Recommender Systems Cheng et al. 2016 — Wide & Deep Learning for Recommender Systems
8 Neural collaborative filtering Neural Collaborative Filtering (NCF) He et al. 2017 — Neural Collaborative Filtering
9 DeepFM for CTR ranking DeepFM for recommendation systems explained Guo et al. 2017 — DeepFM: A Factorization-Machine based Neural Network for CTR Prediction
10 Two-stage candidate generation and ranking Deep Neural Networks for YouTube Recommendations Covington, Adams & Sargin 2016 — Deep Neural Networks for YouTube Recommendations
11 Two-tower retrieval Building Scalable Retrieval System with Two-Tower Models  
12 Self-attentive sequential recommendation Self-Attentive Sequential Recommendation (SASRec) Kang & McAuley 2018 — Self-Attentive Sequential Recommendation

Found a broken link or an unclear step? Report a problem with this track.