Abstract
Universal approximation capability of feedforward neural networks is one of the important theoretical concepts in artificial neural networks. In this study, a type of single-hidden-layer feedforward neural networks is presented. The networks is called feedforward cylindrical approximate identity neural networks. Then, universal approximation capabilities of the networks are investigated in two function spaces. Whereby the notions of cylindrical approximate identity and cylindrical convolution are introduced. The analyses are divided into two cases: In the first case, universal approximation capability of a single-hidden-layer feedforward cylindrical approximate identity neural networks to continuous bivariate functions on the infinite cylinder is investigated. In the latter case, universal approximation capability of the networks is extended to the pth-order Lebesgue integrable bivariate functions on the infinite cylinder.
Similar content being viewed by others
References
Cybenko G.: Approximation by superpositions of a sigmoidal function. Math Control Signals Syst. 2, 303–314 (1989)
Hornik K.: Multilayer feedforward networks are universal approximators. Neural Netw. 2, 359–366 (1989)
Mhaskar H.N.: Approximation properties of a multilayered feedforward artificial neural network. Adv. Comput. Math. 1, 61–80 (1993)
Leshno M., Lin V.Y., Pinkus A., Schocken S.: Multilayer feedforward networks with a nonpolynomial activation function can approximate any function. Neural Netw. 6, 861–867 (1993)
Suzuki S.: Constructive function-approximation by three-layer artificial neural networks. Neural Netw. 11, 1049–1058 (1998)
Pinkus A.: Approximation theory of MLP model in neural networks. Acta Numer. 8, 143–195 (1999)
Castro J.L., Mantas C.J., Benitez J.M.: Neural networks with a continuous squashing function in the output are universal approximators. Neural Netw. 13, 561–563 (2000)
Back A.D., Chen T.: Universal approximation of multiple operators by neural networks. Neural Comput. 14, 2561–2566 (2002)
Liao Y., Fang S.C., Nuttle H.L.W.: Relaxed conditions for radial basis function networks to be universal approximators. Neural Netw. 16, 1019–1028 (2003)
Wang J., Xu Z.: New study on neural networks: the essential order of approximation. Neural Netw. 23, 618–624 (2010)
Arteaga C., Marrero I.: Approximation in weighted p-mean by RBF networks of Delsarte translates. J. Math. Anal. Appl. 414, 450–460 (2014)
Ismailov V.E.: On the approximation by neural networks with bounded number of neurons in hidden layers. J. Math. Anal. Appl. 417, 963–969 (2014)
Costarelli D.: Interpolation by neural network operators activated by ramp functions. J. Math. Anal. Appl. 419, 574–582 (2014)
Cao H., Cao F., Wang D.: Quantum artificial neural networks with applications. Inf. Sci. 290, 1–6 (2015)
Tang Y., Gao H., Kurths J.: Multiobjective identification of controlling areas in neuronal networks. IEEE Trans. Comput. Biol. Bioinf. 10, 708–720 (2013)
Tang Y., Gao H., Lu J., Kurths J.: Pinning distributed synchronization of stochastic dynamical networks: a mixed optimization approach. IEEE Trans. Neural Netw. Learn. Syst. 25, 1804–1815 (2014)
Tang Y., Wang Z., Gao H., Kurths J.: On controllability of neuronal networks with constraints on the average of control gains. IEEE Trans. Cybern. 44, 2670–2681 (2014)
Turchetti C., Conti M., Crippa P., Orcioni S.: On the approximation of stochastic processes by approximate identity neural networks. IEEE Trans. Neural Netw. 9, 1069–1085 (1998)
Panahian Fard, S.; Zainuddin, Z.: SO(2) approximate identity neural networks are universal approximators. 7th International Conference on Neural Computation Theory and Applications, pp. 146–150 (2015)
Panahian Fard, S.; Zainuddin, Z.: Almost everywhere approximation capabilities of double Mellin approximate identity neural networks. Soft Comput. (2015). doi:10.1007/s00500-015-1753-y
Panahian Fard S., Zainuddin Z.: The universal approximation capabilities of double periodic approximate identity neural networks. Soft Comput. 19, 2883–2890 (2015)
Panahian Fard S., Zainuddin Z.: Analyses for L p[a, b]-norm approximation capability of flexible approximate identity neural networks. Neural Comput. Appl. 24, 45–50 (2014)
Panahian Fard S., Zainuddin Z.: On the universal approximation capability of flexible approximate identity neural networks. Lecture Notes in Electrical Engineering 236, 201–207 (2013)
Zainuddin Z., Panahian Fard S.: Double approximate identity neural networks universal approximation in real Lebesgue spaces. Lect. Notes Comput. Sci. 7663, 409–415 (2012)
Panahian Fard S., Zainuddin Z.: The Universal Approximation Capability of Double Flexible Approximate Identity Neural Networks. Lect. Notes Electr. Eng. 277, 125–133 (2014)
Panahian Fard S., Zainuddin Z.: The universal approximation capabilities of Mellin approximate identity neural networks. Lect. Notes Comput. Sci. 7951, 205–213 (2013)
Panahian Fard, S.; Zainuddin, Z.: The universal approximation capabilities of 2\({\pi}\)-periodic approximate identity neural networks. The 2013 International Conference on Information Science and Cloud Computing, IEEE, pp. 793–798 (2014)
Zainuddin, Z.; Panahian Fard, S.: Spherical approximate identity neural networks are universal approximators. The 10th International Conference on Natural Computation, IEEE, pp. 72–76 (2014)
Panahian Fard S., Zainuddin Z.: Toroidal approximate identity neural networks are universal approximators. Lect. Notes Comput. Sci. 8834, 135–142 (2014)
Wade J.: Ceasà àro summability of fourier orthogonal expansions on cylinder. J. Math. Anal. Appl. 402, 446–459 (2013)
Arorey A., Facchi P., Man’ko V.I., Marmo G., Pascazio S., Sudarshan E.G.C.: Radon transform on the cylinder and tomography of a particle on the particle. Phys. Rev. A 76, 012117-1–012117-9 (2007)
Stefano D.M., Marchioro M., Sommariva A.: Polynomial approximation and cubature at approximate Fekete and Leja points of the strip. Appl. Math. Comput. 218(2012), 10617–10629 (2012)
Carnicero J.A., Ausin M.C., Wiper M.P.: Non-parametric copulas for circular-linear and circular-circular data: an application to wind directions. Stoch. Environ. Res. Risk Assess. 27, 1991–2002 (2013)
Fernandez-Duran J.J.: Models for circular–linear and circular–circular data constructed from circular distributions based on nonnegative trigonometric sums. Biometrics 63, 579–585 (2007)
Roy, A.; Pauri S.K.; Roy, U.: SWGMM: a semi-wrapped Gaussian mixture model for clustering of circular–linear data. Pattern Anal. Appl. (2014). doi:10.1007/s10044-014-0418-2
Wu W., Nan D., Li Z., Long J.: Approximation to compact set of functions by feedforward neural networks. Proc. Int. Jt. Conf. Neural Netw. 20, 1222–1225 (2007)
Johnson R.A., Wehrly T.E.: Some angular–linear distribution and related regression models. J. Am. Stat. Assoc. 73, 602–606 (1978)
Wang M.: Extensions of Probability Distributions on Torus, Cylinder and Disc. School of Science and Technology, Kieo University, Tokyo (2013)
Lebedev V.: An Introduction to Functional Analysis and Computational Mathematics. Brikhauser, Boston (1997)
Jones F.: Lebesgue Integration on Euclidean Space. Jones and Bartlett, Burlington (1993)
Author information
Authors and Affiliations
Corresponding author
Rights and permissions
About this article
Cite this article
Zainuddin, Z., Fard, S. The Universal Approximation Capabilities of Cylindrical Approximate Identity Neural Networks. Arab J Sci Eng 41, 3027–3034 (2016). https://doi.org/10.1007/s13369-016-2067-9
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s13369-016-2067-9