Feature selection, as a preprocessing step to machine learning, has been very effective in reducing dimensionality, removing irrelevant data, increasing learning accuracy, and imp...
We briefly survey several privacy compromises in published datasets, some historical and some on paper. An inspection of these suggests that the problem lies with the nature of the...
Abstract. In this paper, we present a systematic evaluation of a hybrid approach of combined rule-based filtering and machine learning to Dutch coreference resolution. Through the...
Developing middleware services for dynamic distributed systems, e.g., ad-hoc networks, is a challenging task given that such services deal with dynamically changing membership and...
Kishori M. Konwar, Peter M. Musial, Nicolas C. Nic...
Imitation Learning, while applied successfully on many large real-world problems, is typically addressed as a standard supervised learning problem, where it is assumed the trainin...