Abstract Identifier attributes--very high-dimensional categorical attributes such as particular product ids or people's names--rarely are incorporated in statistical modeling....
We present a novel approach to solving Quantified Boolean Formulas (QBF) that combines a search-based QBF solver with machine learning techniques. We show how classification met...
Partial cognates are pairs of words in two languages that have the same meaning in some, but not all contexts. Detecting the actual meaning of a partial cognate in context can be ...
This paper introduces a novel machine learning model called multiple instance ranking (MIRank) that enables ranking to be performed in a multiple instance learning setting. The mo...
Charles Bergeron, Jed Zaretzki, Curt M. Breneman, ...
One of the most important challenges in supervised learning is how to evaluate the quality of the models evolved by different machine learning techniques. Up to now, we have relied...