The students will:
- familiarize themselves with the concepts of Recommender Systems,
- understand the challenges involved,
- be able to “recommend” appropriate techniques when faced with a recommendation task, and
- acquire hands-on experience implementing existing methods and evaluating them over real datasets.
- Introduction
- Collaborative Filtering (CF)
- Model-based CF -- Matrix Factorization
- Content-based Recommenders
- Evaluation Methods
- Sequence-aware Recommenders
- Special Topics (e.g., Ethics, Group Recommenders, Social Recommenders)
This course is an overview of the general research area of Recommender Systems. The goal of these systems is to address the information overload problem (multitude of choices) people face in everyday life. Examples include selecting news articles to read, a movie to watch, a travel destination, friends to connect with, a restaurant to dine, buying a product.
The course will introduce the basic concepts, that is, users, items, preferences, explicit/implicit feedback, and proceed to explain important tasks, such as modeling a user’s preferences and an item’s attractiveness, collecting feedback from users, predicting the degree of interest of a user for an item, evaluating effectiveness. For these tasks the course will overview the most important approaches taken, and discuss the state-of-the-art. Towards the end of the course, certain advanced specialized topics, recently being investigated by the research community, will be discussed.
The students will be asked to implement simple approaches using real-life datasets, and work on a real-case task (such as an ACM RecSys Challenge).
<p>The course involves programming assingments and a project, all in groups of 3-5 students, and a final written exam.</p>
<p>- 4 Programming Assignments in Python using Jupyter Notebooks, where you fill in the missing code. (20% of total grade)</p>
<p>- 1 Project in any programming language. (30% of total grade)</p>
<p><span>- The written exam is open book with open questions and some T/F statements. (50% of total grade)</span></p>
- no course prerequisites
- background in Machine Learning, Information Retrieval, E-Commerce is welcome but not required
- all necessary concepts are introduced in course
- content on slides alone suffice