In this course, you will learn how Big Tech (Facebook, TikTok, Amazon, Netflix, YouTube, etc.) develops content/product recommendation systems to provide customized experiences, increase engagement, and drive up customer satisfaction. We explore content-based and collaborative filtering paradigms, architectures using statistical methods, deep learning, CNNs, autoencoders, RNNs, Transformers, GANs, and deep RL. We’ll also delve into scoring, re-ranking, evaluation, deployment, ethics, decision-making psychology, and adversarial attacks. For each topic, we’ll cover definitions, reference papers, explore classical methods, look at current research, and list open questions. Lying at the intersection of machine learning and business, this course will be application-focused while prioritizing mathematical/technical rigor.
Pre-requisites (not enforced, but recommended): CS 61A, EECS 16A/Math 54
Methods of Instruction:
- Lectures, interactive & conceptual homework assignments (with coding portions)
- Group work highly recommended (group size 2-4 people). Groups can submit assignments for a shared grade
- 80 min online lecture to go over context, available models, and homework assignments
- 50 min online office hours for conceptual questions and help with homework assignments
- Industry simulations where students will be asked to make recommendation system design decisions
Attendance Policy:
We allow a maximum of 2 unexcused absences to the lecture; if you aren't able to make it to lecture, please email us in advance with a reason for why you can't make it. During each lecture, we provide a simple quiz that makes sure that you are following along with the course content; the quiz is very easy, but you must be present at lecture to take it. The grading policy for the quizzes and other assignments are available on the course website.
No day(s) left until application deadline!
| Name | Download Link | ||
|---|---|---|---|