Keynote Speaker: Shankar Kalyanaraman (Facebook)

Bio: Shankar Kalyanaraman is a research scientist in the Core Data Science team at Facebook. His interests are in the intersection of recommender systems and AI safety. He received a PhD in theoretical computer science from California Institute of Technology in 2009 and an undergraduate degree in CS from IIT Madras in 2001.

Title: Measuring and mitigating long-term effects of recommender systems: A framework and a call to action

Abstract: Modern recommender systems rely on learning users' preferences through the course of interactions with them. Such interactions can occasionally lead to outcomes that are misaligned with the user's original intent. We introduce a framework for modeling and identifying how recommender systems can lead to such effects and propose strategies to mitigate them.

Program Schedule ( September 25, 3:00 PM to 7:00 PM CEST)


  • 3:00 PM - 3:10 PM Introduction and welcome

  • 3:15 PM - 4:15 PM Keynote (Facebook)

  • 4:15 PM - 4:35 PM Dmitri Goldenberg, Javier Albert and Guy Tsype. Optimization Levers for Promotions Personalization Under Limited Budget (long)

  • 4:35 PM - 4:50 PM Alain Starke, Christoph Trattner,Hedda Bakken, Martin Johannessen and Vegard Solberg. The Cholesterol Factor: Balancing Accuracy and Health in Recipe Recommendation Through a Nutrient-Specific Metric (short)

  • 4:50 PM - 5:05 PM Sinan Seymen, Himan Abdollahpouri and Edward Carl Malthouse. A unified optimization toolbox for solving popularity bias, fairness, and diversity in recommender systems (short)


BREAK


  • 5:30 PM - 5:50 PM Tiago Cunha, Ioannis Partalas and Phong Nguyen. Juggler: Multi-Stakeholder Ranking with Meta-Learning (long)

  • 5:50 PM - 6:05 PM Sasha Stoikov and Hongyi Wen. Evaluating Music Recommendations with Binary Feedback for Multiple Stakeholders (short)

  • 6:05 PM - 6:20 PM Arnault Pachot, Adélaïde Albouy-Kissi, Benjamin Albouy-Kissi and Frédéric Chausse. Multiobjective recommendation for sustainable production systems (short)

  • 6:20 PM - 6:35 PM Blagoj Mitrevski, Milena Filipovic, Diego Antognini, Emma Lejal Glaude, Boi Faltings and Claudiu Musat. Momentum-based Gradient Methods in Multi-Objective Recommendation (short)

  • 6:35 PM - 7 PM Panel and Discussion