A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2019; you can also visit the original URL.
The file type is
Lecture Notes in Computer Science
As the main result we show that every reasonably sized standard network of radial basis function (RBF) neurons has VC dimension (W log k), where W is the number of parameters and k the number of nodes. ... We establish superlinear lower bounds on the Vapnik-Chervonenkis (VC) dimension of neural networks with one hidden layer and local receptive eld neurons. ... Acknowledgment This work has been supported in part by the ESPRIT Working Group in Neural and Computational Learning II, NeuroCOLT2, No. 27150. ...doi:10.1007/3-540-44581-1_2 fatcat:ji2r6x2s4fabrmu7lckky6fpv4
Neural nets with superlinear VC-dimension. Neural Comp. 6, 877-884. VC Dimension of Parameterized Function Classes 1053 Powell, M. J. D. 1987. ... Certain two- and three-layer feedforward neural networks with tanh(x) activation function are known to have VC dimension at least 2(W log W). ...doi:10.1162/neco.19184.108.40.2060 fatcat:hv5ybwmotbfrtmvm77ebgzirvi
Lecture Notes in Computer Science
We establish versions of Descartes' rule of signs for radial basis function (RBF) neural networks. ... This result contrasts with previous work showing that RBF neural networks with two and more input nodes have superlinear VC dimension. ... In this contribution we show that versions of Descartes' rule of signs can be established for radial basis function (RBF) neural networks. ...doi:10.1007/3-540-36169-3_26 fatcat:x5yclcltivez5fywlky2ezvbui
As to lower bounds, we construct product unit networks of fixed depth with superlinear VC dimension. ... In particular, we derive upper and lower bounds on the Vapnik-Chervonenkis (VC) dimension and the pseudodimension for various types of networks with multiplicative units. ... Some of the results have been presented at the NeuroCOLT Workshop "New Perspectives in the Theory of Neural Nets" in Graz, Austria, on May 3, 2000. ...doi:10.1162/08997660252741121 pmid:11802913 fatcat:qivw7gldfvf33ndge4s66robcy
Topics discussed include nearest neighbor, kernel, and histogram methods, Vapnik-Chervonenkis theory, and neural networks. ... Radial Basis Function Classifiers: These classifiers have a "neural" flavor and base their decision upon the sign of functions of the form where is a kernel function (such as or ), is an integer, and the ... However, typical radial basis function classifiers choose , and tune the rest of the parameters according to some criterion. ...doi:10.1109/18.720536 fatcat:pboyft5ze5gwphln5bpglatbam
INDEX TERMS Extreme learning machine, function optimization, Harris's hawk optimization, stock prediction, support vector machine, synthetic aperture radar target recognition. ... An Improved exponential decreasing inertia weight-particle swarm optimization algorithm  is combined with a radial basis function neural network and is applied to perform the air quality index prediction ... In this paper, the penalty parameter C and kernel function parameter γ in radial basis function (RBF) are two parameters of SVM that need to be optimized. ...doi:10.1109/access.2020.2985596 fatcat:arih4dvnz5etrmat4vvcffslru
(English summary) 2004j:68156 Schmitt, Michael Radial basis function neural networks have superlinear VC dimension. ... (English summary) (see 2004j:68005) — RBF neural networks and Descartes’ rule of signs. ...
Tutorials were given by Christoph von der Malsburg ("The binding problem of neural networks"), Wulfram Gerstner ("Models of spiking neurons"), Angus Mac-intyre ("The VC-dimension of neural networks"), ... computing and learning on biological and artificial neural networks. ... For supervised learning the described incremental models can be coupled with the radial basis function (RBF) approach. This leads to incremental RBF networks. ...fatcat:hw6zmihh3jestkyhrjgswc7igu
Transportation and mobility networks, Internet, mobile phone networks, power grids, social and contact networks, neural networks, are all examples where space is relevant and where topology alone does ... An important consequence of space on networks is that there is a cost associated to the length of edges which in turn has dramatic effects on the topological structure of these networks. ... connections exist between regions with no direct structural connections, demonstrating that the structural and functional properties of neural networks are entangled in a complex way and that future studies ...doi:10.1016/j.physrep.2010.11.002 fatcat:mp77spbw2jewvijlttu7y3bktq
(Bastien Fernandez) 2004m:37154 37L60 (37A25, 37C30) Schmitt, Michael Radial basis function neural networks have superlinear VC dimension. ... (see 2004j:68005) 68T05 (68Q32) — RBF neural networks and Descartes’ rule of signs. (English summary) Algorithmic learning theory, 321-335, Lecture Notes in Comput. ...
Processing of continuous speech by a hierarchical neural network. ... The speech vectors are then quantized into a determined number of categories using a self-organizing neural network. ... This is the basis of the IPA problem. ...doi:10.1121/1.410770 fatcat:ioiiov6bmjdi7kiflait5dhdfe
Processing of continuous speech by a hierarchical neural network. ... The speech vectors are then quantized into a determined number of categories using a self-organizing neural network. ... This is the basis of the IPA problem. ...doi:10.1121/1.413700 pmid:7608403 fatcat:m3nunzs4wfflhfl27n4shdn2n4
Processing of continuous speech by a hierarchical neural network. ... The speech vectors are then quantized into a determined number of categories using a self-organizing neural network. ... This is the basis of the IPA problem. ...doi:10.1121/1.411137 fatcat:ajwj7bozxzg6fd27v4xe57drgm
Book of Abstracts
Acknowledgments Having used equation of mass transfer equation and numerical methods, we improved this technique and determined distribution coefficients, heights of theoretical plates and influence of ... The intensity of the lower-energy VIS emission shows the superlinear dependence on the Bi 3+ content. ... Good properties which both of those two materials have in common are strain in compression, tensile strength, good accuracy and precision, very solid dimensional stability, fine elastic recovery and minimal ...doi:10.21175/rad.abstr.book.2021.17.7 fatcat:ltnqghblr5e6blnq6zimmsdyxu