Abstract. A crucial problem in machine learning is to choose an appropriate representation of data, in a way that emphasizes the relations we are interested in. In many cases this ...
Documents formatted in eXtensible Markup Language (XML) are becoming increasingly available in collections of various document types. In this paper, we present an approach for the...
Massih-Reza Amini, Anastasios Tombros, Nicolas Usu...
The notion of embedding a class of dichotomies in a class of linear half spaces is central to the support vector machines paradigm. We examine the question of determining the mini...
In many application domains there is a large amount of unlabeled data but only a very limited amount of labeled training data. One general approach that has been explored for util...
Avrim Blum, John D. Lafferty, Mugizi Robert Rweban...
Aiming to clarify the convergence or divergence conditions for Learning Classifier System (LCS), this paper explores: (1) an extreme condition where the reinforcement process of ...