- BEHW89.A. Blumer, A. Ehrenfeucht, D. ttaussler, and M. K. Warmuth. Learnability and the Vapnik-Chervonenkis dimeusion. J. ACM, 36(4):929-965, 1989. Google ScholarDigital Library
- DH73.R, O. Duds and P. E. Hart. Paller~ Classzficaizon and Scene Analysis. Wiley, 1973.Google Scholar
- DK95.E. Dichterman and R. Khardon. A tight bound for the VC dimension of k-term DNF. Private communication.Google Scholar
- Kha79.L.G. Khachiyan. A polynomial algorithm in linear programming (in Russian). Dok'- lady Akadcmlz Nauk SSSR, 244:1093-1096, 1979. (English translation: Smqet Mathematics Doklady 20:191-194, 1979.)Google Scholar
- KW94.J. Kivinen and M. K. Warmuth. Exponentiated gradient versus gradient descent for linear predictors. Report UCSC-CRL- 94-16, University of California, Santa Cruz, June 1994. Google ScholarDigital Library
- Lit88.N. Littlestone. Learning quickly when irrelevant attributes abound: A new linearthreshold algorithm. Mach. Learmng, 2:285-318, 1988. Google ScholarDigital Library
- Lit89.N. Little~tone. M~ttake Bounds and Logamthm~c Linear-threshold Learning Algorithms. PhD thesis, Report UCSC-CRL- 89-11, University of California Santa Cruz, 1989, Google ScholarDigital Library
- Lit91.N. Littlestone, Redundant noisy attributes, attribute errors, and linear threshold learning using Winnow. In Proc. dth Workshop on Comput. Learning Theory, pages 147- 156. Morgan Kaufinann, 1991. Google ScholarDigital Library
- Lit95.N. Littlestone. Comparing several linearthreshold learning algorithms on tasks involving superfluous attributes. In Proc. 12th I, ler'natzonal Machine Learning Conference (ML-95), July 1995.Google Scholar
- LLW91.N. Littlestone, P. M. Long, and M. K. Warmuth. On-line learning of linear functions. In Proc. 23rd ACM Symposzum on Theory of Computing, pages 465 475, 1991. Google ScholarDigital Library
- MT94.W. Maass and G. Turgm. How fast can a threshold gate learn. In Computatwnal gear, z,g Theory and Natural Learmng Systema, Volume I, pages 381-414. MIT Press, 1994, Google ScholarDigital Library
- Ros58.F. Rosenblatt. The perceptron: A probabilistic model for information storage and organization in the brain. Psych. Rev., 65:386-407, 1958. (Reprinted in Neurocomputing (MIT Press, 1988).). Google ScholarDigital Library
- SST91.It. Sompolinsky, H. S. Seung, and N. Tishby. Learning curves in large neural networks. In Proc. dth Workshop on Cornpitt. ~ear, m9 Theory, pages 112-127. Morgan Kaufrnann, 1991. Google ScholarDigital Library
- Vai89.P M. Vaidya. A new algorithm for minimizing convex functions over convex sets. In Proc. 30lh Symposzum on Foundatwns of Computer Sczcnce, pages 338 343. IEEE Cornputer Society, 1989.Google ScholarDigital Library
- Val84.L.G. Valiant. A theory of the learnable. Commu,. ACM, 27(11):1134-1142, 1984. Google ScholarDigital Library
- VC71.V N. Vapnik and A. Y. Chervonenkis. On the uniform convergence of relative frequencies of events to their probabilities. Theory of Probab. and ~ts Apphcations, 16(2):264- 280, 1971.Google Scholar
- WRB93.T.L.H. Watkin, A. Rau, and M. Biehl. The statistical mechanics of learning a rule. Rev. Mod. Phys., 65:499-5,56, 1993.Google ScholarCross Ref
Index Terms
- The perceptron algorithm vs. Winnow: linear vs. logarithmic mistake bounds when few input variables are relevant
Recommendations
Perceptron, Winnow, and PAC Learning
We analyze the performance of the widely studied Perceptron and Winnow algorithms for learning linear threshold functions under Valiant's probably approximately correct (PAC) model of concept learning. We show that under the uniform distribution on ...
PAC Analogues of Perceptron and Winnow Via Boosting the Margin
We describe a novel family of PAC model algorithms for learning linear threshold functions. The new algorithms work by boosting a simple weak learner and exhibit sample complexity bounds remarkably similar to those of known online algorithms such as ...
Comments