Recommender Systems Workshop

Course Information

• Time: Tuesday 16:00-18:00
• Location: Check Point 002
• Moodle: Workshop on Recommender Systems (Moodle)
• Previous courses: Recommender Systems Workshop
• Data sets: Public Data Sets
• Course ID: 0368-3565

Course Staff

• Dr. Rubi Boim
• Stav Koren (TA) - stavk AT mail.tau.ac.il

Course Goal

Ever wondered how Instagram curates your feed or selects posts for the Explore page? Or how Netflix recommends shows tailored just for you? The same technology powers Spotify, YouTube, and even Tinder.

This technology is called Recommender Systems — a specialized branch of Machine Learning designed to predict user behavior.
Whether it’s making a purchase, clicking a post, playing a song, or swiping right, these systems analyze massive amounts of data to anticipate your next move.
They are also a major driver of revenue, generating billions of dollars for companies by increasing engagement, retention, and sales.

The goal of this workshop is for you to build your own recommender system. The exact domain, use case and focus are entirely up to you!

Course Format

The project will be done by groups of 4 students.

• Fundamentals of Recommender Systems
  We will explore the foundational concepts, algorithms, and mathematical principles (e.g., matrix factorization, similarity measures) that drive recommender systems, along with broader considerations such as scalability and system design.
  
  
• Team Project Proposals
  Students will form small teams (4 members) and propose a specific project related to recommender systems.
  (Each proposal must be approved before moving forward)
  
  
• Implementation and Presentation
  Teams will develop and test their projects throughout the semester, including a mid-semester presentation to share progress and get feedback. At the end of the semester, teams will submit a written report and deliver a final presentation detailing their approach, results, and lessons learned.

Course Requirements

• Significant self-learning is expected
  While the project work will be intensive, students will find it both fun and rewarding, ensuring an engaging and enjoyable learning experience.

Course Schedule and Slides

* tentative due to Roaring Lion operation

Reference Papers, Classes and Others

• Trends in Recommendation & Personalization at Netflix
• X For You Feed Algorithm
• Twitter's recommendation algorithm
• Instagram's Explore recommender system
• Scaling the Instagram Explore recommendations system
• New is always better

Recommender systems frameworks

* this is a "random" list. Please let us know which ones are good :)

• Building a MovieLens Recommender System
• Building a Movie Recommendation Engine in Python using Scikit-Learn
• Build a Book Recommender System with Python: A Machine Learning Project Tutorial
• List of Recommender Systems
• A Python scikit for recommender systems
• LightFM
• Implicit