Mathematical model of architecture and learning processes of artificial neural networks

• Autorzy: Bielecki A.
• Języki publikacji: EN

Abstrakty

EN

A mathematical model of architecture and learning processes of multilayer artificial neural netwoks is discussed in the paper. Dynamical systems theory is used to describe the learning precess of networks consisting of linear, weakly nonlinear and nonlinear neurons. Conjugacy between a gradient dynamical system with a constant time step and a cascade generated by its Euler method theorem is applied as well.

Słowa kluczowe

EN

artificial neural networks; neuron; learning process; topological conjugacy; gradient dynamical system; Euler method

• Wydawca: Politechnika Gdańska
• Czasopismo: TASK Quarterly : scientific bulletin of Academic Computer Centre in Gdansk
• Rocznik: 2003
• Tom: Vol. 7, No 1
• Strony: 93--114
• Opis fizyczny: Bibliogr. 47 poz.

Twórcy

autor

Bielecki A.

• Institute of Computer Science, Jagiellonian University, Nawojki 11, 30-072 Cracow, Poland

Bibliografia

• [1] Hertz J, Krogh A and Palmer R G 1991 Introduction to the Theory of Neural Computation,Addison-Welsey Publishing Company, Massachusetts
• [2] Korbicz J, Obuchowicz A and Uciński D 1994 Artificial Neural Networks – Backgrounds andApplications, Akademicka Oficyna Wydawnicza PLJ, Warsaw (in Polish)
• [3] Osowski S 1996 Neural Networks – the Algorithmic Approach, WNT, Warsaw (in Polish)
• [4] Rutkowska D, Piliński M and Rutkowski L 1997 Neural Networks, Genetic Algorithms andFuzzy Systems, PWN, Warsaw-Lodz (in Polish)
• [5] Żurada J, Barski M and Jędruch W 1996 Artificial Neural Networks, PWN, Warsaw(in Polish)
• [6] Tadeusiewicz R 1993 Neural Networks, Akademicka Oficyna Wydawnicza, Warsaw (in Polish)
• [7] Podolak I T 1998 IEEE Trans. on Systems, Men and Cybern., Part B: Cybernetics28(6) 876
• [8] Owens A J and Filkin D L 1989 Int. Joint Conf. on Neural NetworksII, Washington,New York: IEEE, pp. 381–386
• [9] Kloeden P E and Lorenz J 1986 SIAM J. Numer. Anal.23986
• [10] Alouges F and Debussche A 1991 Numer. Funct. Anal. Optimiz.12253
• [11] Beyn W J 1987 SIAM J. Numer. Anal.24103
• [12] Beyn W J and Lorenz J 1987 Numer. Funct. Anal. Optimiz.9381
• [13] Feˇckan M 1991 Proc. Amer. Math. Soc.111585
• [14] Feˇckan M 1991 Proc. Amer. Math. Soc.1131105
• [15] Mrozek M and Rybakowski K P 1992 J. Dyn. Diff. Eq.457
• [16] Srzednicki R 1994 J. Diff. Eq.111283
• [17] Garay B 1996 J. Diff. Eq. App.267
• [18] Feˇckan M 1992 Math. Slovaca42(1) 123
• [19] Garay B 1993 Acta Math. Univ. Comenianae62245
• [20] Bielecki A 2000 Ann. Pol. Mat.7337
• [21] Garay B 1994 Acta Math. Univ. Comenianae6325
• [22] Garay B 1996 J. Numer. Math.4449
• [23] Garay B 1998 IMA J. Numer. Anal.1877
• [24] Dudek-Dyduch E 1986 Cybernetics of Neuro-like Systems, Zakład Nar. im. Ossolińskich,PAN Publ., Wroclaw (in Polish
• [25] Bielecki A 2001 Opuscula Mathematica20249
• [26] Bielecki A 2001 Nonlinear Analysis: Real World Applications2249
• [27] Campbell S A 1998 Differential Equations with Applications to Biology(Ruan S, Wolkow-icz G S K and Wu J, Eds.)2165
• [28] Deville Y 1996 Signal Processing51(3) 229
• [29] Gelenbe E 1990Neural Computation2(2) 239
• [30] Hoppensteadt F C and Izhikevich E M 1997 Weakly Connected Neural Networks, Springer
• [31] Tanaka K 1996 IEEE Trans. on Neural Networks7(3) 629
• [32] Solla S 1995 Phys. Rev. Lett.744337
• [33] Solla S and Levin E 1986 Complex Systems2625
• [34] Saad D and Solla S 1996 Advances in Neural Information Processing Systems8302
• [35] Duch W and Jankowski N 1999 Neural Computing Surveys2163
• [36] Bielecki A 1998 Gradient Dynamical Systems and a Training Process of Multilayer NeuralNetworks, PhD Thesis, Faculty of Mathematics and Physics, Jagiellonian University, Cracow(in Polish)
• [37] Podolak I T 1998 Comput. Phys. Commun.117(1–2) 62
• [38] Baldi P and Hornik K 1995 IEEE Trans. on Neural Networks6(4) 837
• [39] Demidowicz B 1972 Mathematical Stability Theory, WNT, Warsaw (in Polish)
• [40] Wiggins S 1990 Introduction to Applied Nonlinear Dynamical Systems and Chaos, Springer-Verlag, New York
• [41] Voyevodin V V 1980 Linear Algebra, Mir Publishers, Moscow
• [42] Palis J and de Melo W 1982 Geometric Theory of Dynamical Systems, Springer Verlag,New York
• [43] Irwin M C 1987 Smooth Dynamical Systems, Academic Press Ltd., London
• [44] M ̈uller B and Reinhardt J 1990 Neural Networks, Springer Verlag, New York
• [45] Tadeusiewicz R 1994 Problems of Biocybernetics, PWN, Warsaw (in Polish)
• [46] Jabłoński D and Bielecki A 2000 Proc. 6thNational Conf. on Application of Mathematics inBiology and Medicine, Zawoja, Poland, pp. 57–60
• [47] Ma ̃n ́e R 1988 Publ. Mathematiques IHES66161