《02 - Learning at the Knowledge Level》.pdf

Machine Learning 1 : 287-316, 1986 © 1986 Kluwer Academic Publishers, Boston - Manufactured in The Netherlands Learning at the Knowledge Level THOMAS G. DIETTERICH (DIETTERICH%OREGON-STATE@CSNET-RELAY) Department of Computer Science, Oregon State University, Corvallis, OR 97331, U.S.A. (Received January 10, 1986) (Revised April 15, 1986) Key words: inductive learning , deductive learning , knowledge level , learning theory Abstract. When Newell introduced the concept of the knowledge level as a useful level of description for computer systems, he focused on the representation of knowledge . This paper applies the knowledge level notion to the problem of knowledge acquisition. Two interesting issues arise . First , some existing machine learning programs appear to be completely static when viewed at the knowledge level . These programs improve their performance without changing their knowledge. Second, the behavior of some other machine learning programs cannot be predicted or described at the knowledge level . These programs take unjustified inductive leaps. The first programs are called symbol level learning (SLL) programs; the second, nondeductive knowledge level learning (NKLL) programs. The paper analyze s both of these classes of learning programs and speculates on the possibility of developing coherent theories of each. A theory of symbol level learning is sketched, and some reasons are presented for believing that a theory of NKL L will be difficult to obtain . 1. Introduction In his AAAI Presidents Address, Allen Newell (1981) defined a level of computer system description called the knowledge level. As with other