Please use this identifier to cite or link to this item:
http://hdl.handle.net/1942/4598
Title: | Training binary perceptron by clipping | Authors: | SCHIETSE, Jan BOUTEN, Marcus VAN DEN BROECK, Christian |
Issue Date: | 1995 | Source: | Europhysics letters, 32(3). p. 279-284 | Abstract: | A teacher perceptron T with N binary components provides the classification of a set of p randomly chosen training examples. Several algorithms are available that use this information to select a student perceptron J with continuous components Ji. The purpose is to maximize the overlap R = T·J/(|T| |J|), or to minimize the corresponding generalization error ε = (1/π) arccos R. In view of the binary nature of the components of the teacher, one might expect that a lower error can be achieved by working with the clipped version of the student vector, namely the vector with components sign (Ji). It turns out that this is not always the case. In this letter we calculate the overlap for a vector with components f(Ji), where f can be any odd function of its argument, as a function of the overlap R. We show that the optimal choice of f is a hyperbolic tangent f(x) = th ((R/(1 - R2)x)). The corresponding generalization error can go to zero exponentially fast in a2, for a large (a = p/N). | Document URI: | http://hdl.handle.net/1942/4598 | DOI: | 10.1209/0295-5075/32/3/015 | Type: | Journal Contribution |
Appears in Collections: | Research publications |
Show full item record
SCOPUSTM
Citations
10
checked on Sep 2, 2020
WEB OF SCIENCETM
Citations
10
checked on Jul 13, 2024
Page view(s)
40
checked on Nov 7, 2023
Google ScholarTM
Check
Altmetric
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.