Abstract
In this paper, we investigate the complexity of exact-learning classes of permutation groups and linear spaces over finite fields. These concept classes can be seen as special subclasses of the concept class of circuits. It turns out that in Angluin's exact-learning model these concept classes are learnable with equivalence queries but not learnable with membership queries, indeed not even exact learnable in a probabilistic sense.
More importantly, we raise the question whether the entire power of equivalence queries is needed to learn these concept classes. In order to answer this question we introduce the notion of a teaching assistant. Intuitively speaking, the exact learning model comprising of the teacher and learner, is enhanced to include an intermediate teaching assistant with which the learning algorithm communicates. Building on ideas from complexity theory, we quantify the complexity of teaching assistants for learning permutation groups and linear spaces. As our main theorems, we show:
-
The representation class Sym of permutation groups are exactly learnable with an LWPP-assistant.
-
The class of linear spaces over finite fields are exactly learnable with an SPP-assistant.
Since LWPP and SPP are low complexity classes, this means that these concept classes are of low complexity. As a negative result, we show that if 3-CNFs are exactly learnable with an LWPP-assistant (SPP-assistant), then NP \(\subseteq\)LWPP (respectively NP \(\subseteq\)SPP).
Preview
Unable to display preview. Download preview PDF.
References
D. Angluin. Queries and concept learning. Machine Learning, 2:319–342, 1988.
D. Angluin. Negative results for equivalence queries. Machine Learning, 5:121–150, 1990.
A. Blummer, A. Ehrenfeucht, D. Haussler and M. Warmuth. Learnability and the Vapnik-Chervonenkis dimension. Journal of the ACM, 36(4):929–965, 1989.
J. L. Balcázar, J. DÃaz, R. Gavaldà and O. Watanabe. A note on the query complexity of learning DFA. Proceedings of the Third Workshop on Algorithmic Learning Theory, 53–62, 1992.
S. Fenner, L. Fortnow, S. Kurtz. Gap-definable counting classes. Proc. 6th Structure in Complexity Theory Conference, 30–42, 1991.
M. Furst, J. E. Hopcroft and E. Luks. Polynomial time algorithms for permutation groups. Proc. 21st IEEE Symposium of Foundations of Computer Science, 36–45, 1980.
R. Gavaldá. On the power of equivalence queries. EUROCOLT, 193–203, 1993.
M. Hall. The Theory of Groups. Macmillan, New York, 1959.
L. Hellerstein, K. Pillaipakkamnatt, V. Raghavan and D. Vilkins. How many queries are needed to learn? Proc. of the 27th Symposium on the Theory of Computing, 190–199, 1995.
K. Hoffmann and R. Kunz. Linear Algebra. Prentice Hall Inc., 1971.
J. Köbler, U. Schöning, J. Torán. Graph isomorphism is low for PP. Journal of Computational Complexity, 2:301–310, 1992.
L. Valiant. A theory of the learnable. Communications of the ACM, 27(11):1134–1142, 1984.
O. Watanabe. A formal study of learning via queries. 17th International Colloquium on Automata, Languages and Programming, LNCS 443:139–152, 1990.
O. Watanabe and R. Gavaldá. Structural analysis of polynomial time query learnability. Mathematical Systems Theory, 27:231–256, 1994.
Author information
Authors and Affiliations
Editor information
Rights and permissions
Copyright information
© 1996 Springer-Verlag Berlin Heidelberg
About this paper
Cite this paper
Arvind, V., Vinodchandran, N.V. (1996). The complexity of exactly learning algebraic concepts. In: Arikawa, S., Sharma, A.K. (eds) Algorithmic Learning Theory. ALT 1996. Lecture Notes in Computer Science, vol 1160. Springer, Berlin, Heidelberg. https://doi.org/10.1007/3-540-61863-5_38
Download citation
DOI: https://doi.org/10.1007/3-540-61863-5_38
Published:
Publisher Name: Springer, Berlin, Heidelberg
Print ISBN: 978-3-540-61863-8
Online ISBN: 978-3-540-70719-6
eBook Packages: Springer Book Archive