• A Tutorial on Feature Interpretation in Recommender Systems
    by Zhaocheng Du, Chuhan Wu, Qinglin Jia, Jieming Zhu and Xu Chen


    Data-driven techniques have greatly empowered recommender systems in different scenarios. However, many mainstream algorithms rely on black-box models, making them difficult to interpret, debug, and evolve. Therefore, effectively and efficiently interpreting the behaviors and impacts of features in different stages of recommendation pipelines is essential in industrial recommender systems to master a clear picture of the features they use and bring new insights to system improvement and product design. In this tutorial, we present a systematic overview of feature interpretation technologies in the recommendation field from various aspects including algorithms, applications, and challenges. We first provide a systematic taxonomy of previous feature interpretation methods based on their interpretation perspectives, then introduce the experience and lessons of feature interpretation in large-scale and real-time industrial recommender systems. Finally, we summarize several remaining theoretical and practical challenges in feature interpretation and present corresponding future directions to help feature interpretation better empower recommender systems. From this tutorial, the RecSys community can obtain insights into the methodology and real-world applications of feature interpretation to make more transparent, targeted, and intelligent system optimization.

  • Computational Methods for Designing Human-Centered Recommender Systems: A Case Study Approach Intersecting Visual Arts and Healthcare
    by Bereket Yilma


    Designing and developing modern-day Recommender Systems (RecSys) is a multi-disciplinary effort that benefits from advancements obtained in different computer science fields, particularly Machine learning, Information retrieval and human-computer interaction (HCI)[2]. To harness the full potential of RecSys engines, professionals and researchers must equip themselves with a holistic understanding of not only the computational methods enabling the design and development of these systems but also the know-how to ensure human aspects are the centre of the design. Hence, this course approaches Recsys from a human-centred perspective, looking at the interface and algorithm studies that advance understanding of how system designs can be tailored to users’ objectives and needs while taking into account external factors such as commercialization. By participating in this course, attendees will acquire a comprehensive understanding of the computational methods to design human-centric RecSys, encompassing fundamental concepts, advanced algorithms, and practical implementation with a more emphasis on putting humans at the heart of the design process. This course takes a case study approach to RecSys from an HCI perspective intersecting visual arts with a healthcare application.

  • Conducting Recommender Systems User Studies Using POPROX
    by Robin Burke, Joseph Konstan and Michael Ekstrand


    The Platform for OPen Recommendation and Online eXperimentation (POPROX) is a new resource to allow RecSys researchers to conduct online user research without having to develop all of the necessary infrastructure and recruit users. Our first domain is personalized news recommendations – POPROX 1.0 provides a daily newsletter (with content from the Associated Press) to users who have already consented to participate in research, along with interfaces and protocols to support researchers in conducting studies that assign subsets of users to various experimental algorithms and/or interfaces.

    The purpose of this tutorial is to introduce the platform and its capabilities to prospective research users while walking through the implementation of a sample experiment so that researchers can proceed to propose and carry out experiments on the POPROX platform.

  • Conducting User Experiments in Recommender Systems
    by Bart Knijnenburg and Edward Malthouse


    Traditionally, the field of recommender systems has evaluated the fruits of its labor using metrics of algorithmic accuracy and precision. In recent years, however, researchers have come to realize that the goal of a recommender system extends well beyond accurate predictions; its primary real-world purpose is to provide personalized help in discovering relevant content or items. This realization has caused prominent recommender systems researchers to call for a broadening of the scope of research beyond algorithms and beyond accuracy- or precision-based evaluation.
    Despite these calls, surprisingly little recommender systems research focuses on preference elicitation, the presentation of recommendations, and/or other aspects of the “Human-Recommender Interaction”. Similarly, very few researchers evaluate their recommenders in online user experiments with subjective and experience-based metrics.

    While our papers, book chapters, and past tutorials on the user experience of recommender systems have been instrumental in raising awareness regarding these topics, we believe that a lack of more in-depth training in user-centric design and evaluation methods remains an important reason for the relative lack of user-centric recommender systems research. We therefore believe that a tutorial on the user experience of recommender systems is both timely and important. It will provide practical training in conducting user experiments and statistical analysis of the results of such experiments, thereby helping researchers and practitioners improve the user experience of the recommender systems they develop. In the long term, this will trigger more scientific user-centric work that can grow our knowledge on how certain recommender system aspects influence the user experience.

  • Deep Recommendation using Graphs
    by Panagiotis Symeonidis


    The tutorial will offer a rich blend of theory and practice. It is important for the RecSys community because it describes graph-based algorithms to model the user preferences and to address the similarity search and recommendation problem. This problem affects our everyday experience while searching for knowledge on a topic with challenging issues such as scalability, noise, and sparsity. We can deal with all the challenges by applying graph-based methods. We will provide a detailed step-by-step analysis, by using an integrated toy example with notebooks using python, which runs throughout main parts of the tutorial, and helps the audience to clearly understand the differences among different graph-based methods.

  • Economics of Recommender Systems
    by Emilio Calvano, Giacomo Calzolari and Vincenzo Denicolo


    We propose a tutorial on the economics of recommender systems, examining their impacts on product markets, individual choices, and economic actors’ incentives. The goal is to foster interdisciplinary collaboration and promote knowledge exchange. The tutorial covers theoretical foundations, empirical analysis, experimental research, and policy considerations. Key areas include causal effects, surplus generation, bias, market power, and policy implications. Targeting a diverse audience, it offers insights from leading scholars in the field.

Tutorials Chairs

  • Federica Cena, University of Turin, Italy
  • Christoph Trattner, MediaFutures, University of Bergen, Norway
  • Martijn Willemsen, Eindhoven University of Technology & JADS, The Netherlands