Volume 37, Number 6, February III 1997
|Page(s)||427 - 432|
|Section||Cross-disciplinary physics and related areas of science and technology|
|Published online||01 September 2002|
Noise robustness in multilayer neural networks
Limburgs Universitair Centrum - B-3590 Diepenbeek, Belgium
2 Institut für Theoretische Physik, Universität Würzburg, Am Hubland, D-97074 Würzburg, Germany
3 Instituto de Física, Universidade de São Paulo, CP 66318, 05389-970 São Paulo, SP, Brazil
Accepted: 17 January 1997
The training of multilayered neural networks in the presence of different types of noise is studied. We consider the learning of realizable rules in nonoverlapping architectures. Achieving optimal generalization depends on the knowledge of the noise level, however its misestimation may lead to partial or complete loss of the generalization ability. We demonstrate this effect in the framework of online learning and present the results in terms of noise robustness phase diagrams. While for additive (weight) noise the robustness properties depend on the architecture and size of the networks, this is not so for multiplicative (output) noise. In this case we find a universal behaviour independent of the machine size for both the tree parity and committee machines.
PACS: 87.10.+e – General, theoretical, and mathematical biophysics (including logic of biosystems, quantum biology, and relevant aspects of thermodynamics, information theory, cybernetics, and bionics) / 02.50.-r – Probability theory, stochastic processes, and statistics / 07.05.Mh – Neural networks, fuzzy logic, artificial intelligence
© EDP Sciences, 1997
Current usage metrics show cumulative count of Article Views (full-text article views including HTML views, PDF and ePub downloads, according to the available data) and Abstracts Views on Vision4Press platform.
Data correspond to usage on the plateform after 2015. The current usage metrics is available 48-96 hours after online publication and is updated daily on week days.
Initial download of the metrics may take a while.