Abstract
We introduce a new fault-tolerant model of algorithmic learning using an equivalence oracle and anincomplete membership oracle, in which the answers to a random subset of the learner's membership queries may be missing. We demonstrate that, with high probability, it is still possible to learn monotone DNF formulas in polynomial time, provided that the fraction of missing answers is bounded by some constant less than one. Even when half the membership queries are expected to yield no information, our algorithm will exactly identifym-term,n-variable monotone DNF formulas with an expectedO(mn 2) queries. The same task has been shown to require exponential time using equivalence queries alone. We extend the algorithm to handle some one-sided errors, and discuss several other possible error models. It is hoped that this work may lead to a better understanding of the power of membership queries and the effects of faulty teachers on query models of concept learning.
Article PDF
Similar content being viewed by others
References
Angluin, D. (1987). Learning regular sets from queries and counterexamples.Information and Computation, 75, 87–106.
Angluin, D. (1988). Queries and concept learning.Machine Learning, 2, 319–342.
Angluin, D. (1990). Negative results for equivalence queries.Machine Learning, 5, 121–150.
Angluin, D., Frazier, M., & Pitt, L. (1992). Learning conjunctions of Horn clauses.Machine Learning, 9, 147–164.
Angluin, D., Hellerstein, L., & Karpinski, M. (1993). Learning read-once formulas with queries.JACM, 40, 185–210.
Angluin, D., & Kharitonov, M. (1991). When won't membership queries help? InProceedings of the Twenty-Third Annual ACM Symposium on Theory of Computing, (pp. 444–454). New Orleans, LA: ACM Press.
Angluin, D., & Laird, P. (1988). Learning from noisy examples.Machine Learning, 2, 343–370.
Angluin, D., & Slonim, D. (1991). Learning monotone DNF with an incomplete membership oracle. InProceedings of the Fourth Annual Workshop on Computational Learning Theory, (pp. 139–146). Santa Cruz, CA: Morgan Kaufmann.
Goldman, S., Kearns, M., & Schapire, R. (1990). Exact identification of circuits using fixed points of amplification functions. InProceedings of the Thirty-First Annual Symposium on Foundations of Computer Science, (pp. 193–202). St. Louis, MO: IEEE Computer Society Press.
Hancock, T. (1990). Identifying μ-formula decision trees with queries. InProceedings of the Third Annual Workshop on Computational Learning Theory, (pp. 23–37). Rochester, NY: Morgan Kaufmann.
Ishizaka, H. (1990). Polynomial time learnability of simple deterministic languages.Machine Learning, 5, 151–164.
Karp, R.M. (1991). Probabilistic recurrence relations. InProceedings of the Twenty Third Annual ACM Symposium on Theory of Computing, (pp. 190–197). New Orleans, LA: ACM Press.
Kearns, M. (1990).The Computational Complexity of Machine Learning. Cambridge, MA: MIT Press. (Also, Doctoral dissertation, Harvard University, 1989.)
Kearns, M., & Li, M. (1988). Learning in the presence of malicious errors. InProceedings of the Twentieth Annual ACM Symposium on Theory of Computing, (pp. 267–280). Chicago, IL: ACM Press.
Kearns, M., Li, M., Pitt, L., & Valiant, L.G. (1987). On the learnability of Boolean formulae. InProceedings of the Nineteenth Annual ACM Symposium on Theory of Computing, (pp. 285–295). New York, NY: ACM Press.
Kearns, M., & Valiant, L. (1989). Cryptographic limitations on learning boolean formulae and finite automata. InProceedings of the Twenty-First Annual ACM Symposium on Theory of Computing, (pp. 433–444). Seattle, WA: ACM Press.
Laird, P. (1987).Learning from Good Data and Bad. Doctoral dissertation, Department of Computer Science, Yale University, New Haven, CT. (Published asLearning from Good and Bad Data, Kluwer Academic Publishers, 1988.)
Littlestone, N. (1988). Learning quickly when irrelevant attributes abound: A new linear-threshold algorithm.Machine Learning, 2, 285–318.
Maass, W. (1991). On-line learning with an oblivious environment and the power of randomization. InProceedings of the Fourth Annual Workshop on Computational Learning Theory, (pp. 167–175). Santa Cruz, CA: Morgan Kaufmann.
Raghavan, V., & Schach, S. (1990). Learning switch configurations. InProceedings of Third Annual Workshop on Computational Learning Theory, (pp. 38–51). Rochester, NY: Morgan Kaufmann.
Sakakibara, Y. (1991). On learning from queries and counterexamples in the presence of noise.Information Processing Letters, 37, 279–284.
Shackelford, G. & Volper, D. (1988). Learning k-DNF with noise in the attributes. InProceedings of the 1988 Workshop on Computational Learning Theory, (pp. 97–103). Cambridge, MA: Morgan Kaufmann.
Sloan, R. (1988). Types of noise in data for concept learning. InProceedings of the 1988 Workshop on Computational Learning Theory, (pp. 91–96). Cambridge, MA: Morgan Kaufmann.
Sloan, R. (1989).Computational Learning Theory: New Models and Algorithms. Doctoral dissertation, MIT Laboratory for Computer Science.
Valiant, L. (1984). A theory of the learnable.CACM, 27, 1134–1142.
Valiant, L. (1985). Learning disjunctions of conjunctions. InProceedings of the 9th IJCAI, (pp. 560–566). Los Angeles, CA: Morgan Kaufmann.
Author information
Authors and Affiliations
Rights and permissions
About this article
Cite this article
Angluin, D., Slonim, D.K. Randomly fallible teachers: Learning monotone DNF with an incomplete membership oracle. Mach Learn 14, 7–26 (1994). https://doi.org/10.1007/BF00993160
Issue Date:
DOI: https://doi.org/10.1007/BF00993160