This paper introduces a novel regularization strategy to address the generalization issues for large-margin classifiers from the Empirical Risk Minimization (ERM) perspective. Fi...
Dwi Sianto Mansjur, Ted S. Wada, Biing-Hwang Juang
To reveal information hiding in link space of bibliographical networks, link analysis has been studied from different perspectives in recent years. In this paper, we address a no...
When data collection is costly and/or takes a significant amount of time, an early prediction of the classifier performance is extremely important for the design of the data minin...
We evaluate a new hybrid language processing approach designed for interactive applications that maintain an interaction with users over multiple turns. Specifically, we describe ...
In multi-agent planning environments, action models for each agent must be given as input. However, creating such action models by hand is difficult and time-consuming, because i...