讲座题目：Improving on Ranking Through Boosting
主 讲 人：Harold Connamacher Assistant professor Department of Electrical Engineering and Computer Science, Case Western Reserve University
Harold Connamacher is an Assistant Professor in the department of Electrical Engineering and Computer Science at the Case Western Reserve University. He received his Ph.D. in Computer Science from the University of Toronto. His general interests are in studying the underlying structure of problems. Specific areas of research include graph theory, constraint satisfaction problems, random structures, and algorithms. This research tends to lie in the border between theoretical computer science, artificial intelligence, and statistical physics.
In machine-learned ranking, we are given a training set that consists of a set of elements, a distribution over those elements, and a partial ordering of the elements. The goal is to learn a ranking function that will correctly rank novel pairs of elements. Now, suppose that as part of the training set we are given a set of ranking functions that perform poorly relative to the given partial order. The new goal is to combine the poor rankers into an ensemble function that will do a good job of ranking novel pairs of elements. Boosting is the process by which we combine weak learners into a strong ensemble learner. This talk gives an overview of how we use boosting with ranking functions, describes new results, and introduces important open questions.