Menu Close

Keynotes

Christine Bauer | Three views of a secret: Embracing multiple perspectives in the evaluation of information retrieval and recommender systems

As users are increasingly confronted with information and choice overload, we need the ‘right’ information, at the ‘right’ time, in the ‘right’ place, in the ‘right’ way, to the ‘right’ person. Information retrieval and recommender systems are effective means to address this goal. When optimizing and evaluating such systems, we often disregard that a ‘typical’ user is not the only stakeholder interested in a well-functioning system. Beyond ignoring the needs of specific stakeholders, this eventually leads to a malfunctioning system for anyone. In this talk, I will demonstrate that we need to consider the demands of the various stakeholders and provide insights into how we can embrace those needs when evaluating our systems.

Christine Bauer is EXDIGIT Professor for Interactive Intelligent Systems at the Department of Artificial Intelligence and Human Interfaces (AIHI) at the Paris Lodron University Salzburg, Austria. Her research centers on interactive intelligent systems. In recent years, she worked on context-aware recommender systems in the music and media domains. The core interests in her research activities are fairness and multi-method evaluations. She has authored more than 100 papers and holds several best paper awards and several awards for her reviewing activities. She is on the Editorial Board of ACM Transactions on Recommender Systems (TORS) and co-organizes the Workshop series “Perspectives on the Evaluation of Recommender Systems (PERSPECTIVES)”. Further information can be found at https://christinebauer.eu

Sean Macavaney | Re-Thinking Re-Ranking

Re-ranking systems take a “cascading” approach, wherein an initial candidate pool of documents are ranked and filtered to produce a final result list. This approach exhibits a fundamental relevance misalignment problem: the most relevant documents may be filtered out by a prior stage as insufficiently relevant, ultimately reducing recall and limiting the potential effectiveness. In this talk, I challenge the cascading paradigm by proposing methods that efficiently pull in additional potentially-relevant documents during the re-ranking process, using the long-standing Cluster Hypothesis. I demonstrate that these methods can improve the efficiency and effectiveness of both bi-encoder and cross-encoder retrieval models at various operational points. Cascading is dead, long live re-ranking!

Sean is a Lecturer in Machine Learning at the University of Glasgow and a member of the Terrier Team. His research primarily focuses on effective and efficient neural retrieval. He completed his PhD at Georgetown University in 2021, where he was a member of the IR Lab and an ARCS Endowed Scholar. He was a co-recipient of the ECIR 2023 Best Short Paper award.