Five Best Paper Awards

"Best Student Paper Award" at ACM SIGKDD Conference: F. Radlinski and T. Joachims, 'Query Chains: Learning to Rank from Implicit Feedback'. Summary: A method to learn improved retrieval functions from how users reformulate queries and how they click.

"Distinguished Student Paper Award" (one out of four papers) at ICML: Alexandru Niculescu-Mizil and Rich Caruana, 'Predicting Good Probabilities With Supervised Learning'. Summary: Training classifiers to predict good probabilities, not only class assignments.

"Distinguished Student Paper Award" (one out of four papers) at ICML: T. Finley and T. Joachims, 'Supervised Clustering with Support Vector Machines'. Summary: Training a clustering algorithm to produce desirable clusterings based on examples of what kinds of clusterings the user wants.

"Best Paper Award" at ICML: T. Joachims, 'A Support Vector Method for Multivariate Performance Measures'.
Summary: A Support Vector algorithm that can optimize a large class of performance measures (e.g. F1-score, PRBEP), unlike a normal SVM that can only optimize error rate.

"Best Research Paper Award" at KDD 2005: J. Leskovec, J. Kleinberg, C. Faloutsos. Graphs over Time: Densification Laws, Shrinking Diameters and Possible Explanations. Eleventh ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, 2005.