Geodesic
Absence of bottlenecks in a neural network determines its generic functional properties
Doklady Rossijskoj akademii nauk. Matematika, informatika, processy upravleniâ, Tome 490 (2020), pp. 74-77.

Voir la notice de l'article provenant de la source Math-Net.Ru

It is proved that an artificial neural network with smooth activation functions and without bottlenecks is a Morse function for almost all, with respect to the Lebesgue measure, sets of weights.
Keywords: neural network, Morse function. 
@article{DANMA_2020_490_a16,
     author = {S. V. Kurochkin},
     title = {Absence of bottlenecks in a neural network determines its generic functional properties},
     journal = {Doklady Rossijskoj akademii nauk. Matematika, informatika, processy upravleni\^a},
     pages = {74--77},
     publisher = {mathdoc},
     volume = {490},
     year = {2020},
     language = {ru},
     url = {http://geodesic.mathdoc.fr/item/DANMA_2020_490_a16/}
}
TY  - JOUR
AU  - S. V. Kurochkin
TI  - Absence of bottlenecks in a neural network determines its generic functional properties
JO  - Doklady Rossijskoj akademii nauk. Matematika, informatika, processy upravleniâ
PY  - 2020
SP  - 74
EP  - 77
VL  - 490
PB  - mathdoc
UR  - http://geodesic.mathdoc.fr/item/DANMA_2020_490_a16/
LA  - ru
ID  - DANMA_2020_490_a16
ER  - 
%0 Journal Article
%A S. V. Kurochkin
%T Absence of bottlenecks in a neural network determines its generic functional properties
%J Doklady Rossijskoj akademii nauk. Matematika, informatika, processy upravleniâ
%D 2020
%P 74-77
%V 490
%I mathdoc
%U http://geodesic.mathdoc.fr/item/DANMA_2020_490_a16/
%G ru
%F DANMA_2020_490_a16
S. V. Kurochkin. Absence of bottlenecks in a neural network determines its generic functional properties. Doklady Rossijskoj akademii nauk. Matematika, informatika, processy upravleniâ, Tome 490 (2020), pp. 74-77. http://geodesic.mathdoc.fr/item/DANMA_2020_490_a16/

[1] Cybenko G.V., “Approximation by Superpositions of a Sigmoidal function”, Mathematics of Control Signals and Systems, 2:4 (1989), 303–314 | DOI | MR | Zbl

[2] Pinkus A., “Approximation Theory of the MLP Model in Neural Networks”, Acta Numerica, 8 (1999), 143–195 | DOI | MR | Zbl

[3] Zhuravlev Yu.I., Rudakov K.V., “Ob algebraicheskoi korrektsii protsedur obrabotki (preobrazovaniya) informatsii”, Problemy prikladnoi matematiki i informatiki, Nauka, M., 1987, 187–198 | MR

[4] Kurochkin S.V., “Raspoznavanie gomotopicheskogo tipa ob'ekta s pomoschyu differentsialno-topologicheskikh invariantov approksimiruyuschego otobrazheniya”, Kompyuternaya optika, 43:4 (2019), 611–617 | DOI

[5] Postnikov M.M., Vvedenie v teoriyu Morsa, Nauka, M., 1971 | MR

[6] Arnold V., “Smooth functions statistics”, Funct. Anal. Other Math, 2006, no. 1, 111–118 | DOI | MR | Zbl

[7] Gudfellou Ya., Bendzhio I., Kurvill A., Glubokoe obuchenie, DMK Press, M., 2017

[8] Le C., “A note on Optimization with Morse Polynomials”, Commun. Korean Math. Soc, 33:2 (2018), 671–676 | DOI | MR | Zbl

[9] Banyaga A., Hurtubise D., Lectures on Morse Homology, Kluwer Texts Math. Sci., 29, Kluwer Acad. Publ., Dordrecht, 2004 | DOI | MR | Zbl

[10] Prasolov V.V., Elementy kombinatornoi i differentsialnoi topologii, MTsNMO, M., 2014

[11] Baksalary J.K., Kala R., “The Matrix Equation AX-YB = C”, Linear Algebra and Its Applications, 30 (1979), 41–43 | DOI | MR

[12] Prasolov V.V., Zadachi i teoremy lineinoi algebry, MTsNMO, M., 2015

[13] Nicolaescu L., An Invitation to Morse Theory, Springer, 2011 | MR | Zbl

[14] Carlsson G., “Topology and Data”, Bulletin of the American Mathematical Society, 46:2 (2009), 255–308 | DOI | MR | Zbl