MACHINE LEARNING: A Critique of Research Efforts and Suggested Research Strategy
William S. Davis, David B. Murrell · Purdue e-Pubs (Purdue University System) · 1988
Learning is constructing or modifying representations of what is being experienced. and then using those representations to perform be[ter a second time on repetition of the same task, or another task drawn from the same population. Machine Learning is concerned with developing computationallheories of such learning, as well as constructing learning systems. This document is three-fold in nature: 1) a brief Qverview of machine leaming; 2) a critique of some accomplishments in machine learning research; 3) a proposed s[J'ategy for directing furore machine learning research. The aulhors' overall view of machine learning is that, while much of the research results have provided a gocxi foundation in learning theories and taxonomies of learning strategies, a lTu.e learning element for the machine is clearly lacking. This viewpoint will be evident from the critiques. The proposed strategy for research will be [0 narrow the focus of research to the two main problems in the field: the lack of full integration of learning strategies, and the lack of real creatiVity/intelligence in learning.