Abstract
In the supervised learning paradigm, each training example is described by a set of attributes and a class label. However, in many learning situations, class labels are not given; instead, they are replaced by another set of attributes. We call this type of learning cooperative unsupervised learning. The task of cooperative unsupervised learning is to (re)construct class labels consistent with multiple sources of attributes. We design an algorithm, called AutoLabel, that learns class labels from unlabeled training examples described by two sets of attributes. We test AutoLabel on several artificial and real-world datasets, and show that it constructs classification labels accurately. Our learning paradigm removes the fundamental assumption of provision of class labels in supervised learning, and gives a new perspective to unsupervised learning.
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
Reference
de Sa, V. (1994a). Learning classification with unlabeled data. In Cowan, J., Tesauro, G., & Alspector, J. (Eds.), Advances in Neural Information Processing Systems 6.
de Sa, V. (1994b). Minimizing disagreement for self-supervised classification. In Mozer, M., Smolensky, P., Touretzky, D., & Weigend, A. (Eds.), Proceedings of the 1993 Connectionist Models Summer School.
Fisher, D. (1987). Knowledge acquisition via incremental conceptual clustering. Machine Learning, 2, 139–172.
Fisher, D., & Langley, P. (1990). The structure and formation of natural categories. In Bower, G. (Ed.), The Psychology of Learning and Motivation, Vol. 26. Academic Press.
Genarri, J., Langley, P., & Fisher, D. (1989). Models of incremental concept formation. Artificial Intelligence, 40, 11–61.
Lapedes, A. S., Steeg, E. W., &: Farber, R. M. (1995). Use of adaptive networks to define highly predictable protein secondary-structure classes. Machine Learning, 21, 103–124.
Lu, S., & Chen, K. (1987). A machine learning approach to the automatic synthesis of mechanistic knowledge for engineering decision-making. Artificial Intelligence for Engineering Design, Analysis, and Manufacturing, 1, 109–118.
Medin, D. (1983). Structural principles of categorization. In Tighe, T., & Shepp, B. (Eds.), Perception, cognition, and development. Hillsdale, NJ: Erlbaum.
Murphy, P. M., & Aha, D. W. (1992). UCI Repository of machine learning databases [Machine-readable data repository]. Irvine, CA: University of California, Department of Information and Computer Science.
Quinlan, J. (1993). C4-5: Programs for Machine Learning. Morgan Kaufmann: San Mateo, CA.
Reich, Y. (1992). Ecobweb: Preliminary user's manual. Tech. rep., Department of Civil Engineering, Carnegie Mellon University.
Reich, Y., & Fenves, S. (1991). The formation and use of abstract concepts in design. In Fisher, D., Pazzani, M., & Langley, P. (Eds.), Concept Formation: Knowledge and Experience in Unsupervised Learning. Morgan Kaufmann, CA.
Author information
Authors and Affiliations
Editor information
Rights and permissions
Copyright information
© 1996 Springer-Verlag Berlin Heidelberg
About this paper
Cite this paper
Ling, C.X., Wang, H. (1996). Learning classifications from multiple sources of unsupervised data. In: McCalla, G. (eds) Advances in Artifical Intelligence. Canadian AI 1996. Lecture Notes in Computer Science, vol 1081. Springer, Berlin, Heidelberg. https://doi.org/10.1007/3-540-61291-2_59
Download citation
DOI: https://doi.org/10.1007/3-540-61291-2_59
Published:
Publisher Name: Springer, Berlin, Heidelberg
Print ISBN: 978-3-540-61291-9
Online ISBN: 978-3-540-68450-3
eBook Packages: Springer Book Archive