Recommender Systems are becoming an inherent part of today's Internet. They can be found anywhere from e-commerce platforms (eBay, Amazon) to music or movie streaming (Spotify, Netflix), social media (Facebook, Instagram, TikTok), travel platforms (Booking.com, Expedia), and much more. Whether a recommendation is successful or not can rely on multiple objectives such as user satisfaction, business value, and societal issues. In addition, the long-term happiness (along with short-term excitements and delight) of the users is critical for a recommender system to be considered successful. MORS workshop brings together researchers and practitioners to discuss the importance of these aspects of recommender systems and find ways to develop algorithms to build multi-objective recommenders and also evaluation metrics to assess their success. Additional Key Words and Phrases: multi-objective recommendation, Value-aware recommendation
Recommender systems are software tools that are used in a variety of application domains supporting users to find relevant items, products, and services easier. Historically, the main criterion for a successful recommender system was the relevance of the recommended items to the user. In other words, the only objective for the recommendation algorithm was to learn user preferences for diferent items and generate recommendations accordingly. However, real-world recommender systems are well beyond a simple objective and often take into account multiple objectives. Indeed, diferent objectives can be important and should be considered for generating recommendations. These objectives can be either from the users’ perspective or they could come from other stakeholders such as item providers and the ones that could be impacted by the recommendations.
From the users’ perspective, often multiple objectives need to be considered for generating the recommendations. For example, in restaurant recommendations, several factors, such as users’ taste, diet restrictions, the proximity of the restaurant, and even the price could be taken into account. Each of these aspects can be important for the users, some of which, more than the others. Therefore, it is crucial for the recommender system to incorporate all these diferent objectives and aspects into account when recommending some restaurants to the user. Similarly, in the education domain, a student may prefer working on simpler problems to achieve higher scores. However, some struggle is inevitable for the students to learn the new concepts properly. As a result, a problem recommender algorithm should balance the simplicity of the recommended problems and their utility to help students learn more. ∗Copyright 2022 for this paper by its authors. Use permitted under Creative Commons License Attribution 4.0 International (CC BY 4.0). Presented at the MORS workshop held in conjunction with the 16th ACM Conference on Recommender Systems (RecSys), 2022, in Seattle, WA, USA.
The objectives may also come from other stakeholders such as the item providers (those who provide the items to the platform to be recommended) or other stakeholders such as the platform owner or even side stakeholders such as the society. For example, on a music streaming service, the platform may want to balance multiple objectives, some of which are related to the users and some related to the providers (artists) and even the society as a whole. For instance, the platform may want to ensure a certain degree of exposure for diferent artists, so they reach their desired audience and to avoid monopoly by some superstars. The platform may also want to make sure it does not negatively afect the music culture of some smaller countries by over-exposing the users in those countries to some popular western music. These types of objectives and considerations exist in many other domains including social media, transportation, news recommendation, and food recommendation, to name a few.
MORS 2022 Main Theme. In MORS 2022, we focused on group discussions in the research community around
the topics of multi-objective recommender systems. In addition, we had a main "theme" of Long-term Optimization
in Recommender Systems. We believe this aspect of recommendations has been largely overlooked in the research
community, and it is crucial to pay more attention to this aspect. Long-term optimization refers to the fact that the
recommendations given to the users should help achieve long-term satisfaction for the users rather than focusing only
on the short-term metrics (e.g., precision, recall, etc.). The MORS 2022 workshop was a continuation of the discussion of
these topics in prior RecSys workshops including MORS 2021 [
Particularly, the MORS workshop encouraged submissions addressing the following topics of interest: • Recommender systems with multiple objectives • Balancing the long-term impacts of the recommendations and the users’ short-term preferences • Reinforcement Learning for long-term optimization in recommender systems • Evaluation metrics and methodologies for long-term optimization in recommendation • Feedback loops and the impact of recommendations in long term • Value-aware recommendation (profit, value, purpose, etc.) • Recommendation with multiple stakeholders • Conflict handling in multi-stakeholder recommendation • Fairness-aware recommender systems • News recommendation with editorial values • Educational recommender systems with multiple, potentially conflicting, objectives 2
Himan Abdollahpouri is a Research Scientist at Spotify. He was one of the co-chairs of MORS 2021 [
Sherry Sahebi is an assistant professor of Computer Science at the University At Albany – SUNY and the founder of
Personalized AI (PersAI) Lab. Sherry has completed her Ph.D. in Intelligent Systems from University of Pittsburgh in
2016. Since then, she has received several internal (SUNY) and external (NSF) fundings for her projects on recommender
systems and educational data mining. She has been actively involved in the research community, e.g., being a program
chair of the 14th Educational Data Mining conference (EDM’21) and a proceedings chair of the 14th ACM RecSys
(RecSys’20). She has prior experience in successfully co-organizing the Workshop on Intelligent Recommender Systems
by Knowledge Transfer & Learning (RecSysKTL) in RecSys 2017 [
Mehdi Elahi is an Associate Professor at University of Bergen (UiB), Department of the Information Science & Media
Studies (InfoMedia) where he is the co-founder of the Data Analytics and Recommender Systems lab (DARS) in Norway.
He has been involved in the authorship of several grant proposals such as a large-scale grant, recently funded SFI
MediaFutures with a budget of nearly NOK 300 Million, where he will serve as WP Leader for 8 years. He has published
more than 80 peer-reviewed publications in top-tier journals and conferences. His current #citation is 2300+ and his
H-index is 22. Together with big IT companies such as Spotify, XING, and Trivago, he has organized international
challenges in 2017-2019 [
Masoud Mansoury is a postdoctoral researcher at Amsterdam Machine Learning Lab (AMLab) at University of
Amsterdam, Netherlands. He is also a member of Discovery Lab collaborating with Elsevier Company in the area of
recommender systems. Masoud received his PhD in Computer and Information Science from Eindhoven University
of Technology, Netherlands, in 2021. He was one of the co-chair of MORS 2021 workshop [
Babak Loni is a senior Machine Learning Engineer at Meta. Babak has a Ph.D. in Machine Learning and Recommender Systems and a MS.c. in Computer Science, both from Delft University of Technology. He has been organizing RecSysNL meetups and a few RecSys workshops in the past, including the last year’s MORS workshop. Babak has worked in ING, Padora Media, and DPG Media in the past where he built diferent solutions for personalization and recommendations.
Zahra Nazari is a senior research scientist at Spotify working on the design and evaluation of recommender systems
with a focus on cold-start problem, multi-objective recommendations and long-term optimization. Zahra did her Ph.D.
at the University of Southern California with a focus on preference elicitation and human behavior modeling in complex
situations such as negotiations. Her proposed agent won the top prize at the International Automated Negotiating
Agent Competitions in 2017 [
Maria Dimakopoulou is Senior Manager of ML Engineering in the Home Personalization product area at Spotify. She leads the Home Ranking team and the Home Assembly team, which focus on combining cutting-edge advances in ML, causal inference and reinforcement learning with scalable engineering to deliver Spotify Home recommendations that optimize for the long-term satisfaction of the 400+ million Spotify users. Before joining Spotify, Maria was at Netflix, building out the causal recommenders research investment and subsequently starting and leading the Adaptive Experimentation team. Prior to that, she did her PhD on reinforcement learning at Stanford and worked at Google. 3
The following is a list of program committee members.
• Konstantin Bauman (Temple University, USA) • Ludovico Boratto (University of Cagliari, Italy) • Robin Burke (University of Colorado, Boulder, USA) • Dietmar Jannach (Alpen-Adria-Universität Klagenfurt, Austria) • Toshihiro Kamishima (National Institute of Advanced Industrial Science and Technology, Japan) • Mesut Kaya (Aalborg University, Denmark) • Peter Knees (Vienna University of Technology, Austria) • Dominik Kowald (Know-Center, Austria) • Kun Lin (DePaul University, USA) • Yue Shi (Meta, CA, USA)
Annelien Smets (Brussels Centre for Urban Studies, Belgium)
Marko Tkalcic (University of Primorska, Slovenia) • • • 4
The following is the timeline of MORS 2022: • First call for participation: April 15th, 2022 • Paper submission deadline: August 5th, 2022 • Notification of paper acceptance: August 27th, 2022 • Camera-ready version deadline: September 10th, 2022
Workshop (at RecSys 2022): September 23rd, 2022 5
The workshop starts with a keynote by Prof. Dietmar Jannach titled “Multi-Objective Recommendation: Overview
and Challenges”[
Match in Course Recommendation (short)
• [
A Multi-Objective Framework (long)
• [
Systems (poster)
• [
Patricia Grao. Mitigating Targeting Bias in Content Recommendation with Causal Bandits (poster) 6
This work was supported by industry partners and the Research Council of Norway with funding to MediaFutures: Research Centre for Responsible Media Technology and Innovation, through The Centres for Research-based Innovation scheme, project number 309339. It was also partially supported by the United States National Science Foundation under Grant No. 2047500. This project was also funded by Elsevier’s Discovery Lab. 6