Storage capacity of a constructive learning algorithm

DOI10.1088/0305-4470/33/9/301MaRDI QIDQ4495995zbMATH OpenOpenAlexFDO

Authors Arnaud Buhot, Mirta B. Gordon

Publication date 13 August 2000

Published in Journal of Physics A: Mathematical and General (Search for Journal in Brave)

Full work available at URL https://arxiv.org/abs/cond-mat/9909248

Mathematics Subject Classification ID

Learning and adaptive systems in artificial intelligence (68T05) Neural nets applied to problems in time-dependent statistical mechanics (82C32) Neural networks for/in biological studies, artificial life and related topics (92B20)

Abstract: Upper and lower bounds for the typical storage capacity of a constructive algorithm, the Tilinglike Learning Algorithm for the Parity Machine [M. Biehl and M. Opper, Phys. Rev. A {�f 44} 6888 (1991)], are determined in the asymptotic limit of large training set sizes. The properties of a perceptron with threshold, learning a training set of patterns having a biased distribution of targets, needed as an intermediate step in the capacity calculation, are determined analytically. The lower bound for the capacity, determined with a cavity method, is proportional to the number of hidden units. The upper bound, obtained with the hypothesis of replica symmetry, is close to the one predicted by Mitchinson and Durbin [Biol. Cyber. {�f 60} 345 (1989)].

Recommendations

• The statistical mechanics of constructive algorithms
• Storage capacity of a fully-connected parity machine with continuous weights
• Bounds on the learning capacity of some multi-layer networks
• Capacity of the single-layer perceptron and minimal trajectory training algorithms

Cited in

(6)

• Storage capacity of the fully-connected committee machine
• Storage capacity of a fully-connected parity machine with continuous weights
• Capacity of the single-layer perceptron and minimal trajectory training algorithms
• scientific article; zbMATH DE number 2095945 (Why is no real title available?)
• Storage capacity and optimal learning of Potts-model perceptrons by a cavity method
• The statistical mechanics of constructive algorithms

This page was built for publication: Storage capacity of a constructive learning algorithm

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q4495995)