Building Fair and Explainable Recommendation Systems: Theory, Practice, and Immersive Experiences (REMIX)
Organizers
- Raciel Yera Toledo, University of Jaén. Email: ryera@ujaen.es
- Álvaro Labella Romero, University of Jaén. Email: alabella@ujaen.es
- Luis Martínez López, University of Jaén. Email: martin@ujaen.es
Description
Recommender systems (RS) are the main engine that personalizes our digital experiences. With a continuously growing market, the evolution of these systems requires researchers to master not only algorithmic effectiveness, but also the ethical imperatives of artificial intelligence. This 2.5-hour tutorial offers a comprehensive tour of the state of the art in RS.
We will start by laying the theoretical and methodological foundations, exploring fundamental paradigms (collaborative filtering, content-based), evaluation metrics (effectiveness, diversity, coverage) and implementation libraries. Next, we will address one of the greatest current challenges: the lack of transparency and algorithmic bias, with a special focus on Group Recommender Systems (GRS). We will present new directions to ensure that these systems are explainable, fair and neutral with respect to sociodemographic characteristics.
To consolidate the learning, the tutorial includes a guided hands-on session where attendees will implement basic algorithms and fairness/explainability techniques. Finally, the session will culminate with an interactive Virtual Reality demonstration (Meta Quest), allowing attendees to experience first-hand an immersive Top-N recommendation interface in a virtual cinema environment.
