In real-world machine learning problems, it is very common that part of the input feature vector is incomplete: either not available, missing, or corrupted. In this paper, we pres...
Abstract. Smooth boosting algorithms are variants of boosting methods which handle only smooth distributions on the data. They are proved to be noise-tolerant and can be used in th...
Local search is a specialization of the web search that allows users to submit geographically constrained queries. However, one of the challenges for local search engines is to un...