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 analyzes 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 NKLL will be difficult to obtain.
Article PDF
Similar content being viewed by others
References
Angluin, D., & Smith, C.H. (1982). A survey of inductive inference: Theory and methods (Technical Report 250). New Haven, CT: Yale University, Department of Computer Science.
Araya, A. (1984). Learning problem classes by means of experimentation and generalization. Proceedings of the National Conference on Artificial Intellgience (pp. 11–15). Austin, TX: Morgan-Kaufmann.
Bennett, J.S., & Dietterich, T.G. (1986). The test incorporation hypothesis and the weak methods (Technical Report 86–30–4). Corvallis, OR: Oregon State University, Department of Computer Science.
Buchanan, B.G., & Mitchell, T.M. (1978). Model-directed learning of production rules. In D.A. Waterman & F. Hayes-Roth (Eds.), Pattern-directed inference systems (pp. 297–312). New York: Academic Press.
Chase, W., & Simon, H.A. (1973). Perception in chess. Cognitive Psychology, 4, 55–81.
Dietterich, T.G. (1984). Constraint propagation techniques for theory-driven data interpretation (Technical Report STAN-CS-84–1030). Stanford, CA: Stanford University, Department of Computer Science.
Dietterich, T.G., London, R.L., Clarkson, K., & Dromey, G. (1982). Learning and inductive inferenc (Technical Report STAN-CS-82–913). Stanford, CA: Stanford University, Department of Computer Science. In P.R. Cohen & E.A. Feigenbaum (Eds.), The handbook of artificial intelligence (Vol. 3). Los Altos, CA: William Kaufmann.
Dietterich, T.G., & Michalski, R.S. (1981). Inductive learning of structural descriptions: Evaluation criteria and comparative review of selected methods. Artificial Intelligence, 16, 257–294.
Ellman, T. (1985). Generalizing logic circuit designs by analyzing proofs of correctness. Proceedings of the Ninth International Conference on Artificial Intelligence. Los Angeles, CA: Morgan-Kaufmann.
Fikes, R.E., Hart, P.E, & Nilsson, N.J. (1972). Learning and executing generalized robot plans. Artificial Intelligence, 3, 251–288.
Flann, N.S., & Dietterich, T.G. (1985). Exploiting functional vocabularies to learn structural descriptions. Proceedings of the Third International Workshop on Machine Learning (pp. 41–43). New Brunswick, NJ: Rutgers University, Department of Computer Science.
Flann, N.S., & Dietterich, T.G. (1986). Selecting appropriate representations for learning from examples (Technical Report 86–30–5). Corvallis, OR: Oregon State University, Department of Computer Science.
Genesereth, M.R. (1980). Models and metaphors. Proceedings of the National Conference on Artificial Intelligence (pp. 208–211). Stanford, CA: Morgan-Kaufmann.
Gold, E. (1967). Language identification in the limit. Information and Control, 16, 447–474.
Gosling, J. (1983). Algebraic constraints. Doctoral dissertation, Department of Computer Science, Carnegie-Mellon University, Pittsburgh, PA.
Green, C., Luckham, D., Balzer, R., Cheatham, T., & Rich, C. (1983). Report on a knowledge-based software assistant (Technical Report KES.U.83.2). Palo Alto, CA: Kestrel Institute.
Greiner, R., & Genesereth, M.R. (1983). What's new? A semantic definition of novelty. Proceedings of the Eighth International Conference on Artificial Intelligence (pp. 450–454). Karlsruhe, FRG: Morgan-Kaufmann.
Halpern, J.Y., & Moses, Y. (1985). A guide to the modal logics of knowledge and belief: Preliminary draft. Proceedings of the Ninth International Conference on Artificial Intelligence (pp. 480–490). Los Angeles, CA: Morgan-Kaufmann.
Hayes-Roth, F., & McDermott, J. (1978). An interference matching technique for inducing abstractions. Communications of the ACM, 26, 401–410.
Hintikka, J. (1962). Knowledge and belief. Ithaca, NY: Cornell University Press.
Hinton, G.E., Sejnowski, T.J., & Ackley, D.H. (1984). Boltzmann machines: Constraint satisfaction networks that learn (Technical Report CMU-CS-84–119). Pittsburgh, PA: Carnegie-Mellon University, Department of Computer Science.
Keller, R.M. (1983). Learning by re-expressing concepts for efficient recognition. Proceedings of the National Conference on Artificial Intelligence (pp. 182–186). Karlsruhe, FRG: Morgan-Kaufmann.
Kibler, D., & Porter, B. (1983). Episodic learning. Proceedings of the National Conference on Artifical Intelligence (pp. 191–196). Karlsruhe, FRG: Morgan-Kaufmann.
Langley, P. (1983). Learning effective search heuristics. Proceedings of the Eighth International Joint Conference on Artificial Intelligence (pp. 419–421). Karlsruhe, FRG: Morgan-Kaufmann.
Lenat, D.B., & Brown, J.S. (1984). Why AM and EURISKO appear to work. Artificial Intelligence, 23, 269–294.
Lindsay, R.K., Buchanan, B.G., Feigenbaum, E.A., & Lederberg, J. (1980). Applications of artificial intelligence for organic chemistry: The DENDRAL project. New York: McGraw-Hill.
Mahadevan, S. (1985). Verification-based learning: A generalization strategy for inferring problemreduction methods. Proceedings of the Ninth International Conference on Artificial Intelligence (pp. 616–623). Los Angeles, CA: Morgan-Kaufmann.
McCarthy, J. (1958). Programs with common sense. Proceedings of the Symposium on the Mechanization of Thought Processes, Vol. 1 (pp. 77–84). National Physical Laboratory. [Reprinted in M.L. Minsky (Ed.). (1968). Semantic information processing. Cambridge, MA: MIT Press.]
McCarthy, J. (1980). Circumscription—a form of non-monotonic reasoning. Artificial Intelligence, 13, 27–39.
McCarthy, J. (1980). Applications of circumscription to formalizing common sense knowledge. Artificial Intelligence, 28, 89–116.
Michalski, R.S. (1969). On the quasi-minimal solution of the general covering problem. Proceedings of the Fifth International Federation on Automatic Control, 27 (pp. 109–129).
Michalski, R.S., & Chilausky, R.L. (1980). Learning by being told and learning from examples: An experimental comparison of the two methods of knowledge acquisition in the context of developing an expert system for soybean disease diagnosis. International Journal of Policy Analysis and Information Systems, 4, 125–161.
Michalski, R.S., & Larson, J.B. (1978). Selection of most representative training examples and incremental generation of VL1 hypotheses: The underlying methodology and the description of programs ESEL and AQ11 (Technical Report 867). Urbana, IL: University of Illinois, Department of Computer Science.
Mitchell, T.M. (1978). Version spaces: An approach to concept learning (Technical Report STAN-CS-78–711). Stanford, CA: Stanford University, Department of Computer Science.
Mitchell, T.M. (1980). The need for biases in learning generalizations (Technical Report CBM-TR-117). New Brunswick, NJ: Rutgers University, Department of Computer Science.
Mitchell, T.M. (1982). Generalization as search. Artificial Intelligence, 18, 202–226.
Mitchell, T.M. (1983). Learning and problem solving. Proceedings of the Eighth International Joint Conference on Artificial Intelligence (pp. 1139–1151). Karlsruhe, FRG: Morgan-Kaufmann.
Mitchell, T.M., Utgoff, P.E., & Banerji, R. (1983). Learning by experimentation: Acquiring and refining problem-solving heuristics. In R.S. Michalski, J.G. Carbonell, & T.M. Mitchell (Eds.), Machine learning. Los Altos, CA: Morgan-Kaufmann.
Mooney, R., & DeJong, G. (1985). Learning schemata for natural language processing. Proceedings of the Ninth International Joint Conference on Artificial Intelligence (pp. 681–687). Los Angeles, CA: Morgan-Kaufmann.
Mostow, D.J. (1983). Machine transformation of advice into a heuristic search procedure. In R.S. Michalski, J.G. Carbonell, & T.M. Mitchell, (Eds.), Machine learning: An artificial intelligence approach. Los Altos, CA: Morgan-Kaufmann.
Newell, A. (1981). The knowledge level. AI Magazine, 2, 1–20.
Newell, A., & Simon, H.A. (1976). Computer science as empirical inquiry: Symbols and search. Communications of the ACM, 19, 113–126.
Quinlan, J.R. (1983). Learning efficient classification procedures and their application to chess end games. In R.S. Michalski, T.M. Mitchell, & J. G. Carbonell (Eds.), Machine learning: An artificial intelligence approach. Los Altos, CA: Morgan-Kaufmann.
Rosenbloom, P.S., Laird, J.E., Newell, A., Golding, A., & Unruh, A. (1985). Current research on learning in SOAR. Proceedings of the Third International Machine Learning Workshop. New Brunswick, NJ: Rutgers University, Department of Computer Science.
Russell, S. (1985). The compleat guide to MRS (Technical Report KSL-85–12). Stanford, CA: Stanford University, Department of Computer Science.
Samuel, A.L. (1959). Some studies of machine learning using the game of checkers. IBM Journal of Research and Development, 3, 220–229.
Scott, P.D. (1983). Learning: The construction of a posteriori knowledge structures. Proceedings of the National Conference on Artificial Intelligence (pp. 359–363). Karlsruhe, FRG: Morgan-Kaufmann.
Simon, H.A. (1983). Why should machines learn? In R.S. Michalski, T.M. Mitchell, & J.G. Carbonell (Eds.), Machine learning: An artificial intelligence approach. Los Altos, CA: Morgan-Kaufmann.
Stallman, R.M., & Sussman, G.J. (1977). Forward reasoning and dependency-directed backtracking in a system for computer-aided circuit analysis. Artificial Intelligence, 9, 135–196.
Sussman, G.J., & Steele, G.L.Jr. (1980). CONSTRAINTS—A language for expressing almost hierarchical descriptions. Artificial Intelligence, 14, 1–39.
Tappel, S. (1980). Some algorithm design methods. Proceedings of the National Conference on Artificial Intelligence (pp. 64–67). Stanford, CA: Morgan-Kaufmann.
Utgoff, P.E. (1984). Shift of bias for inductive concept learning. Doctoral dissertation, Department of Computer Science, Rutgers University, New Brunswick, NJ.
Utgoff, P.E., & Mitchell, T.M. (1982). Acquisition of appropriate bias for inductive concept learning. Proceedings of the National Conference on Artificial Intelligence (pp. 414–417). Pittsburgh, PA Morgan-Kaufmann.
VanLehn, K. (1983). Felicity conditions for human skill acquisition: Validating an AI-based theory (Technical Report CIS-21). Palo Alto, CA: Xerox Palo Alto Research Center.
Vere, S.A. (1975). Induction of concepts in the predicate calculus. Proceedings of the Fourth International Conference on Artificial Intelligence (pp. 281–287). Tbilisi, USSR: Morgan-Kaufmann.
Author information
Authors and Affiliations
Rights and permissions
About this article
Cite this article
Dietterich, T.G. Learning at the knowledge level. Mach Learn 1, 287–315 (1986). https://doi.org/10.1007/BF00116894
Received:
Revised:
Issue Date:
DOI: https://doi.org/10.1007/BF00116894