Recommender Systems: Perspectives and Research Challenges
Teacher
Giuseppe Sansonetti
Abstract
The amount of content available on the Web is constantly growing and, with it, the difficulties in identifying the relevant information during a search. This experience often causes a sense of overwhelming that can discourage the user well before she can achieve the desired results. The introduction of efficient and reliable Recommender Systems (RSs) by the major web content providers has partly mitigated those issues. Although inexperienced users may not realize it, these techniques are now adopted by most web platforms, thus largely affecting users’ activity: from the choice of books to purchase on Amazon, to the next movie to watch on Netflix.
Program
This course will cover a broad range of topics related to RSs, from an algorithmic as well as an methodological perspective. Furthermore, upcoming and trending topics such as context-aware, group, and affective recommendations approaches will be dealt with.
More specifically, in the first lecture we will introduce RSs, their role in the scientific domain and beyond, and the collaborative filtering technique along with the most advanced issues concerning it. In the second lecture, we will analyze further “core” recommendation techniques such as content-based, utility-based, and knowledge-based approaches. In the third lecture, we will address the concept of context-awareness and how context should impact on recommendations and ratings. Group recommendation will be the topic or the fourth lecture. In particular, we will focus on rank aggregation and balancing techniques. In the last lecture, we will explore some models borrowed from psychology and social sciences and how to exploit them for achieving recommender systems able to take into account the active user’s affective state as well.
Primary Material
Primary source material will be readings in the form of multimedia presentations, research papers and further material provided by the instructor.
Software resources
Credits 3
Length 5 lectures
1. Foundations and Advanced Topics in Collaborative Filtering (Lunedì 15 Luglio 2019, 15:00 – 17:00)
2. Other Main Recommendation Techniques (Mercoledì 17 Luglio 2019, 15:00 – 17:00)
3. Context-Aware Recommendations (Venerdì 19 Luglio 2019, 15:00 – 17:00)
4. Group Recommendations (Lunedì 22 Luglio 2019, 15:00 – 17:00)
5. Affective Recommendations (Mercoledì 24 Luglio 2019, 15:00 – 17:00)