Abstract
This paper describes a novel approach to address the task of learning structural descriptions by incremental representation shifts of learning examples. While in attribute-value representations only one mapping is possible between descriptions, in first order logic based representations there are potentially many mappings. To cope with the intractability of exploring all mappings, classical approaches consider all mappings and then define a restricted hypothesis space. Our approach is to select one particular type of mapping at a time and use it as a basis to define a hypothesis space called Matching Space that may be represented using attribute-value pairs. It appears that characterizing a Matching Space is equivalent to shifting the representation of examples. We provide a proof that Matching Spaces are partially ordered by their size and that there exists a set of Matching Spaces which are, as a whole, equivalent to the initial representation space. Experimental results show the benefits of this approach on a well known learning task.
This is a preview of subscription content, log in via an institution to check access.
Preview
Unable to display preview. Download preview PDF.
Similar content being viewed by others
References
Aha, D. W. (1992). Relating Relational Learning Algorithms. In S. Muggleton (Eds.), Inductive Logic Programming (pp. 233–259). London: Harcourt Brace Jovanovich.
Barrow, H. G., Ambler, A. P., and Burstall, R. M. (1972). Some techniques for recognizing structures in pictures. In S. Watanabe (Eds.), Frontiers of Pattern Recognition (pp. 1–29). New York: Academic.
Cohen, W. (1990). An analysis of Representation Shift In Concept Learning. In B. Porter and R. Mooney (Ed.), Seventh International Conference on Machine Learning, (pp. 104, 112). Austin, Texas.
Dietterich, T., and Michalski, R. (1983). A comparative review of selected methods for learning from examples. In Machine Learning: an Artificial Intelligence Approach San Mateo, CA: Morgan Kaufmann.
Drastal, G., Czako, G., and Raatz, S. (1989). Induction in an abstraction space: A form of constructive induction. In Proceedings of the twelfth International Jointed Conference in Artificial Intelligence (IJCAI), (pp. 708–712). Detroit, Michigan.
Flann, N., and Dietterich, T. (1986). Selecting Appropriate Representations for Learning from Examples. In AAAI, (pp. 460–466). Philadelphia, PA.
Ganascia, J.-G. (1987). CHARADE: A rule System Learning System. In Proceedings of the tenth International Jointed Conference in Artificial Intelligence, (pp. 345–347). Milan, Italy.
Giordana, A., and Saitta, L. (1990). Abstraction: a general framework for learning. In Working notes of the AAAI Workshop on Automated Generation of Approximations and Abstraction, (pp. 245–256). Boston, MA.
Guarino, N., Pribbenow, S., and Vieu, L. (1994). Parts and Wholes: Conceptual Part-Whole Relations and Formal Mereology. In Workshop W2 at the 11th European Conference on Artificial Intelligence, Amsterdam, The Netherlands.
Haussler, D. (1989). Learning Conjunctive Concepts in Structural Domains. Machine Learning, 4, 7–40.
Hayes-Roth, F., and McDermott, J. (1978). An interference matching technique for inducing abstractions. Communications of ACM, 21(5), 401–410.
Kietz, J.-U., and Wrobel, S. (1992). Controlling the Complexity of Learning in Logic through Syntactic and Task-Oriented Models. In S. Muggleton (Eds.), Inductive Logic Programming (pp. 335–359). London: Harcourt Brace Jovanovich.
Köbler, J., Schöning, U., and Torn, J. (1993). The graph-Isomorphism problem: its structural complexity. Boston: Birkhaüser.
Kodratoff, Y., and Ganascia, J.-G. (1986). Improving the generalization step in learning. In J. Carbonell and T. Mitchell (Eds.), Machine Learning: an Artificial Intelligence Approach (pp. 215–244). San Mateo, CA: Morgan Kaufmann.
Korf, R. E. (1985). Depth-First Iterative-Deepening: An optimal Admissible Tree Search. Artificial Intelligence, 27, 97–109.
Lavrac, N., Dzeroski, S., and Grobelnik, M. (1991). Learning Non Recursive Definitions of Relations with Linus. In European Working Session on Learning 91, 482 (pp. 265–281).
Michalski, R. S. (1980). Pattern recognition as rule-guided inductive inference. IEEE Transaction on Pattern Analysis and Machine Intelligence, 2, 349–361.
Michalski, R. S., and Larson, J. B. (1977). Inductive Inference of VL Decision Rules. SIGART Newsletter, ACM(63), 38–44.
Muggleton, S., and Feng, C. (1992). Efficient Induction of logic program. In S. Muggleton (Eds.), Inductive Logic Programming (pp. 281–298). London: Harcourt Brace Jovanovich.
Neri, F., and Saitta, L. (1994). Knowledge Representation in Machine Learning. In F. Bergadano and L. d. Raedt (Ed.), European Conference on Machine Learning, 784. Catania, Italy.
Plotkin, G. (1971). A further note on inductive generalization. In B. Meltzer & D. Michie (Eds.), Machine Intelligence (pp. 101–124). New York: Elsevier.
Quinlan, J.-R. (1990). Learning Logical Definitions from Relations. Machine Learning(5), 239–266.
Shapiro, L. G., and Haralick, R. H. (1980). Structural descriptions and inexact matching. IEEE Transaction Pattern Analysis and Machine Intelligence, 3(5), 504–519.
Sowa, J. F. (1991). Part I: Issues in Knowledge Representation. In J. F. Sowa (Eds.), Principles of Semantic Networks: Exploration in the Representation of Knowledge San Mateo, California: Morgan Kaufmann.
Subramanian, D. (1989). Representational Issues in Machine Learning. In Proceeding of the sixth international Workshop on Machine Learning, (pp. 426–429). California.
Subramanian, D. (1990). A Theory of Justified Reformulations. In D. P. Benjamin (Eds.), Change of Representation and Inductive Bias (pp. 147–167). Boston: Kluwer Academic Publishers.
Winston, M. E., Chaffin, R., and Hermann, D. J. (1987). A taxonomy of Part-Whole relations. Cognitive Science, 11, 417–444.
Zucker, J.-D., and Ganascia, J.-G. (1994). Selective Reformulation of Examples in Concept Learning. In W. Cohen (Ed.), International Conference on Machine Learning, (pp. 352–360). New-Brunswick.
Author information
Authors and Affiliations
Editor information
Rights and permissions
Copyright information
© 1995 Springer-Verlag Berlin Heidelberg
About this paper
Cite this paper
Zucker, JD. (1995). Reformulation of examples in concept learning of structural descriptions. In: Gori, M., Soda, G. (eds) Topics in Artificial Intelligence. AI*IA 1995. Lecture Notes in Computer Science, vol 992. Springer, Berlin, Heidelberg. https://doi.org/10.1007/3-540-60437-5_37
Download citation
DOI: https://doi.org/10.1007/3-540-60437-5_37
Published:
Publisher Name: Springer, Berlin, Heidelberg
Print ISBN: 978-3-540-60437-2
Online ISBN: 978-3-540-47468-5
eBook Packages: Springer Book Archive