Academic Commons

Reports

Concept Learning in a Rich Input Domain: Generalization-Based Memory

Lebowitz, Michael

Automatic concept learning from large amounts of complex input data is an interesting and difficult process. In this paper we discuss how the use of a permanent, generalization-based, memory can serve as an important tool in developing programs that learn in rich input domains. The use of Generalization-Based Memory (GBM) allows programs to determine what concepts to learn, as well as definitions of the concepts. We present in this paper a characterization of our research, describe our use of Generalization-Based Memory in two programs under development at Columbia, UNIMEM and RESEARCHER, and describe how they perform concept evaluation and generalization of complex structural descriptions, problems typical of those we are concerned with.

Subjects

Files

More About This Work

Academic Units
Computer Science
Publisher
Department of Computer Science, Columbia University
Series
Columbia University Computer Science Technical Reports, CUCS-111-84
Published Here
February 17, 2012
Academic Commons provides global access to research and scholarship produced at Columbia University, Barnard College, Teachers College, Union Theological Seminary and Jewish Theological Seminary. Academic Commons is managed by the Columbia University Libraries.