14th Dutch-Belgian Information Retrieval Workshop

27 November, 2015 Amsterdam

Keynotes

Thorsten Joachims

Thorsten Joachims is a Professor in the Department of Computer Science and the Department of Information Science at Cornell University. His research interests center on a synthesis of theory and system building in machine learning, with applications in information access, language technology, and recommendation. His past research focused on support vector machines, text classification, structured output prediction, convex optimization, learning to rank, learning with preferences, and learning from implicit feedback. In 2001, he finished his dissertation advised by Prof. Katharina Morik at the University of Dortmund. From 1994 to 1996 he was a visiting scholar with Prof. Tom Mitchell at Carnegie Mellon University. He is an ACM Fellow, AAAI Fellow, and Humboldt Fellow.

Title: Designing Human Feedback Data for Machine Learning

Abstract: Machine Learning has become one of the key enabling methods for building information access systems. This is evident in search engines, recommender systems, and electronic commerce, while other applications are likely to follow in the near future. In these systems, machine learning is not just happening behind the scenes, but it is increasingly used to directly interact with human users. In fact, much of the Big Data we collect today are the decisions that people make when they use these human-interactive learning systems we built.

In this talk, I argue that for building human-interactive learning system it is crucial to not only design the learning algorithm, but also to design the mechanism for generating the data from the human users. Towards this goal, the talk explores how integrating microeconomic models of human behavior into the learning process leads to new learning models that no longer misrepresent the user as a “labeling subroutine”. This motivates an interesting area for research in information retrieval and machine learning, with connections to rational choice theory, econometrics, and behavioral economics.


Elad Yom-Tov

Elad Yom-Tov is a Principal Researcher at Microsoft Research and a visiting scientist at the Technion, Israel. Before joining Microsoft he was with Yahoo Research, IBM Research, and Rafael. Dr. Yom-Tov studied at Tel-Aviv University and the Technion, Israel. He has published three books, over 60 papers (of which 3 were awarded prizes), and filed more than 30 patents (16 of which have been granted so far). His primary research interests are in using Machine Learning and Information Retrieval to improve health. He is a Senior Member of IEEE and held the title of Master Inventor while at IBM.

Title: Should we optimize search engines for social and personal welfare?

Abstract: Internet search engines are traditionally optimized to rank highest the most relevant results, that is, those results that satisfy the user’s information need. Information need and whether it was met is often judged by users themselves. However, this definition of relevance is subjective to user’s understanding of their information need. In my talk I will show that other notions of relevance are possible, notions that take social or personal welfare into account, and will ask whether we should adopt these alternative relevance measures.

In my talk I will discuss our results in promoting civil discourse through search engine diversity. I will also show that when searching for health information, results deemed relevant by traditional search engines can entrench people in understanding which can lead to unhealthy behaviors or to misunderstanding of one’s medical condition. I will argue that in all these cases, a wider notion of relevance might be required.

I will end with a discussion of the pros and cons of adopting a wider definition of relevance.