Emilia Gómez, Senior Scientist at the European Commission’s Joint Research Center
Chair: Markus Schedl
Date: Tuesday, Sept. 23 – 09:30 – 10:30
Title
Recommender Systems: a European, Science for Policy Perspective
Abstract
The field of recommender systems and its annual conference, RecSys, has traditionally brought together academic researchers and industry experts to discuss advancements in recommendation technologies. However, the increasing use of recommender systems in applications with significant societal impact—such as social media, marketplaces, education, and healthcare—introduces a new perspective that must be considered in the research, development, and deployment of recommender systems: the policy perspective.
In this talk, I will share my personal insights on the challenges that recommender systems pose to society and the related policy issues. I will explore the complexities of working at the intersection of science and policy, highlight some research gaps in the field, and present opportunities for RecSys researchers to contribute to policy-making at different stages of the policy life cycle.
I will provide a European Union-centric view, drawing from my current work at the European Commission’s Joint Research Centre and my research expertise in developing music recommender systems. My keynote will provide a quick overview of the EU Digital Services Act and the AI Act through the lens of recommender system research as connected to algorithmic transparency. Through a series of research projects, I will present some mechanisms to integrate scientific, technical, and policy perspectives in our research field.
Biography
Dr. Emilia Gómez (MSc. Telecommunication Engineering, PhD in Computer Science) joined the JRC in 2018. She leads the Human Behaviour and Machine Intelligence (HUMAINT) team, providing technical and scientific support to EU AI policies, notably the AI Act and the Digital Services Act, as part of the European Centre for Algorithmic Transparency. She is also a guest professor at Universitat Pompeu Fabra in Barcelona and has a long academic experience in the fields of Music Information Retrieval and Human-Centric Machine Learning.
She was the 1st female president of the International Society for Music Information Retrieval, is a member of the OECD One AI expert group, an ELLIS fellow, and her work has been recognized by means of citations and honors, e.g. EUWomen4Future, Red Cross Award to Humanitarian Technologies or ICREA Academia.
Jure Leskovec, Professor of Computer Science at Stanford University
Chair: Pavel Kordik
Date: Wednesday, Sept. 24 – 09:00 – 10:00
Title
Relational Foundation Models: A New Frontier for Predictive AI in Structured Data
Abstract
Foundation Models have transformed how we interact with unstructured data — enabling seamless in-context learning across text, images, and code. Yet, the structured data that drives core decisions in enterprises -— transaction logs, customer journeys, events, time series -— remains locked behind brittle pipelines and handcrafted machine learning models. In this talk, I will introduce Relational Foundation Models (RFMs), a new class of pre-trained models that unlock in-context learning over relational data, just as LLMs did for language.
RFMs model multi-table, heterogeneous graph-structured data and can predict complex outcomes — such as user engagement, purchases, churn, fraud, and recommendations — without per-task supervision, feature engineering, or model training. I will describe the architecture and training objectives of RFMs, which combine table-agnostic embeddings, relational transformers, and SQL-like prompt interfaces. The result is a single general-purpose model that makes accurate, fast predictions across a broad class of tasks, often outperforming traditional supervised pipelines built over months.
We will explore how RFMs reshape the paradigm of predictive AI — from model-building as a craft to model-use as querying — and what this means for the future of recommender systems, classification, regression, and more. I will argue that RFMs are not a replacement for LLMs but their structured-data complement — together forming the foundation for the next generation of enterprise AI.
Biography
Jure Leskovec is Professor of Computer Science at Stanford University. He is affiliated with the Stanford AI Lab, the Machine Learning Group and the Center for Research on Foundation Models. In the past, he served as a Chief Scientist at Pinterest and was an investigator at Chan Zuckerberg BioHub. Most recently, he co-founded AI startup Kumo.AI. Leskovec pioneered the field of Graph Neural Networks and created PyG, the most widely-used graph neural network library. Research from his group has been used by many countries to fight COVID-19 pandemic, and has been incorporated into products at Meta, Pinterest, Uber, YouTube, Amazon, and more. His research received several awards including Microsoft Research Faculty Fellowship in 2011, Okawa Research award in 2012, Alfred P. Sloan Fellowship in 2012, Lagrange Prize in 2015, ICDM Research Contributions Award in 2019, and ACM SIGKDD Innovation award in 2023. His research contributions have spanned social networks, data mining and machine learning, and computational biomedicine with the focus on drug discovery. His work has won 13 best paper awards and 6 10-year test of time awards at premier venues in these research areas. Leskovec received his bachelor’s degree in computer science from University of Ljubljana, Slovenia, PhD in machine learning from Carnegie Mellon University and postdoctoral training at Cornell University.
Xavier “Xavi” Amatriain, VP Product, AI and Compute Enablement at Google
Chair: Maria Bielikova
Date: Thursday Sept. 25, 09:30 – 10:30
Title
Recommending in the Age of AI: How We Got Here and What Comes Next
Abstract
For the recommender systems community, the current AI revolution driven by Generative AI presents both an existential challenge and an unprecedented opportunity. We are on the cusp of realizing a long-held vision: recommending and even generating content for an audience of one. This keynote will provide a historical and future-looking perspective on this pivotal moment. We will begin by reflecting on the Netflix Prize era, recalling how a single competition galvanized our community and embedded machine learning into the core of mainstream products. Drawing on a career that has spanned this evolution —- from Netflix to Quora, LinkedIn, Curai Health, and now Google -— I will share key insights from that journey. We will distinguish between the foundational principles of recommendation that endure and the new rules being written by today’s powerful AI models, illustrated with modern examples from products like YouTube. I will end by discussing the profound implications of emerging technologies like auto-generated personalized media and the complex dynamics of multi-agentic recommender systems.
Biography
Xavier Amatriain (Barcelona 1973) is an AI leader known for his foundational work in recommender systems. He is well-known for leading the teams at Netflix that developed the company’s recommendation engine, a frequently-cited example of translating cutting-edge research into a high-impact product. His career has been built at the intersection of deep research and impactful product development, with subsequent leadership roles driving AI strategy and engineering at Quora and LinkedIn. He also co-founded the health-AI startup Curai Health.
Currently, as VP of Product in the AI and Compute Enablement (ACE) team at Google, Xavier is focused on shaping the core platforms that power Google’s AI-driven products. He holds a Ph.D. in Computer Science from Universitat Pompeu Fabra (Barcelona, Spain) and has authored over 100 research publications on machine learning and related fields.
