The goal of this research is to develop machine-learning approaches for integrating additional resources into the inductive-learning paradigm and to apply these new approaches to numerous real-world problems. Many scientific and industrial problems can be better understood by learning from samples of the task under consideration. For this reason, the machine learning and statistics communities devote considerable research effort in generating inductive-learning algorithms that try to learn the true concept of a task from a set of its examples. Often times, however, once has additional resources readily available, but largely unused, that can significantly improve the concept that these learning algorithms generate. These resources include prior knowledge describing what is currently known about the domain as well as available computer cycles. The focus of this proposal is to extend and apply a successful learning system that consists of utilizing a novel combination of inductive learning (focusing on ensembles, genetic algorithms, and neural networks), anytime learning, and background knowledge. Natural applications of this framework include information filtering, object recognition in digital images, drug design, as well as many other domains traditionally tackled with expert systems. www.cs.umt.edu/CS/FAC/OPITZ/NSF_CAREER.html
