Recommender Systems

Teacher(s)

Vande Kerckhove Corentin;

Language

English

Prerequisites

• Programming in Python
• Elementary probability and statistics
• Mathematics (linear algebra, optimisation, matrix theory)
• Supervised learning (regression/classification) and feature engineering

Main themes

• Recommender system architecture
• Training data for recommender systems: Rating matrices and long-tail distribution of ratings
• Recommender models: collaborative filtering (nearest neighbours, latent factor models), content-based filtering, etc.
• Evaluation of recommender systems

Learning outcomes

At the end of this learning unit, the student is able to :
1	Understand the main currents supporting recommender systems; Understand and describe the main evaluation methods and models used in recommender systems; Apply, on real data, different recommendation techniques, and compare the quality of the results obtained by applying these techniques; Analyze and interpret the results provided by the application of recommendation techniques; Deploy his·her own recommender system.

Content

Nowadays, recommender systems play an ever more important role to propose products or services to consumers. Recommending movies, music, news, books, restaurants, financial services, search terms, or contacts, etc. has become a key asset for many companies. Recommender systems can be based on numerous approaches in existence today. This course covers some of these systems with a focus on recommender systems data, collaborative filtering, matrix factorization, and the evaluation of recommender systems.

Teaching methods

Lectures
Practical assignments, exercises and projects integrated into the course
The lecture is given in English.

Evaluation methods

Continuous evaluation 
Project with oral defense
!!! The course consists in a unique evaluation  (that is, one a mark has been obtained for the course, it holds for the entire academic year, and can NOT been improved later on) !!!
By submitting work for evaluation, you affirm: (i) that it accurately reflects the phenomenon under study, and for this, you must have verified the facts, especially if they are claimed by a generative AI (which you must explicitly mention as a tool used to support your work); (ii) that you have complied with all specific requirements of the task entrusted to you, including requirements for transparency and documentation of the scientific approach implemented. If either of these statements is not true, whether intentionally or due to negligence, you are in violation of your ethical commitment to the knowledge produced in the context of your work, and potentially other aspects of academic integrity, which constitutes an academic offense and will be treated as such.

Other information

Some, or even all, of the lectures might be offered in a hybrid format to enlarge and easier the participation of students from both campuses. The lecturer will provide supplementary information in this sense in due time. 

Online resources

Online ressources are available on Moodle
Lecture name : MLSMM2156 - Systèmes de recommandation
Key : communicated at the first class
Brief introduction: https://tryolabs.com/blog/introduction-to-recommender-systems/
General overview: https://link.springer.com/book/10.1007%2F978-3-319-29659-3

Bibliography

 Aggarwal, Charu C.. “Recommender Systems.” Springer International Publishing (2016).

Faculty or entity

> CLSM