Kudos to grad student Shuo Chen, who received the 2016 KDD Best Student Research Paper Award – Runner Up at the ACM Conference for Knowledge Discovery and Data Mining. Chen is advised by CS Professor Thorsten Joachims. The paper describes a new learning method for representing pairwise preferences, even when these preferences are intransitive and when they depend on context. For example, the method can model that a user may prefer restaurant A over restaurant B for lunch, but that the preference is the other way around for a romantic dinner.

In addition, Joachims’ grad student Tobias Schnabel won the ICTIR Best Presentation Award for his talk at the ACM Conference on the Theory of Information Retrieval. He presented a paper, co-authored by Adith Swaminathan, Peter Frazier, and Thorsten Joachims, on a sampling-based method for comparing the quality of search engines and other AI systems. Unlike the conventional TREC evaluation methodology, the new method is unbiased and requires substantially fewer human relevance judgments to give reliable results.