Abstract
In this letter we present an on-line learning version of the Fokker-Planck machine. The method makes use of a regularized constrained normalized LMS algorithm in order to estimate the time-derivative of the parameter vector of a radial basis function network. The RBF network parametrizes a transition density which satisfies a Fokker-Planck equation, associated to continuous simulated annealing. On-line learning using the constrained normalized LMS method is necessary in order to make the Fokker-Planck machine applicable to large scale nonlinear optimization problems.
Similar content being viewed by others
References
D.L. Alspach and H.W. Sorenson, “Nonlinear Bayesian estimation usingGaussian sum approximations”, IEEE Transactions on Automatic Control, Vol. 17, No. 4, 439–448, 1972.
S.-I. Amari, “Information geometry of the EM and emalgorithms for neural networks”, Neural Networks, Vol. 8, No. 9, 1379–1408, 1995.
J.L. Doob, Stochastic processes, John Wiley & Sons, 1953.
S.B. Gelfand and S.K. Mitter, “Recursive stochastic algorithms for global optimization in ℝd”, SIAM Journal on Control and Optimization, Vol. 29, No. 5, 999–1018, 1991.
S.B. Gelfand and S.K. Mitter, “Metropolistype annealingalgorithms for global optimization in ℝd”, SIAM Journal on Control and Optimization, Vol. 31, No. 1, 111–131, 1993.
I.I. Gihman and A.V. Skorohod, The theory of stochastic processes I, II, III, Springer-Verlag: New York, 1979.
D.E. Goldberg, Genetic algorithms in search, optimization and machine learning,Addison-Wesley: Reading, MA, 1989.
G.H. Golub and C.F. Van Loan, MatrixComputations, Johns Hopkins University Press: Baltimore MD, 1989.
G.C. Goodwin and K.S. Sin, Adaptive filtering, Prediction and Control, Prentice-Hall: Englewood Cliffs, 1984.
S. Haykin, Adaptive Filter Theory, Prentice Hall,1996.
S. Haykin, Neural Networks: a Comprehensive Foundation, MacmillanCollege Publishing Company: Englewood Cliffs, 1994.
H.J. Kushner, “Asymptoticglobal behavior for stochastic approximation and diffusions with slowly decreasing noise effects: global minimization via Monte Carlo”, SIAM Journal on Applied Mathematics, 47, 169–185, 1987.
M. Moonen and J. Vandewalle, “A square root covariance algorithm forconstrained recursive least squares estimation”, Journal of VLSI Signal Processing, 3, 163–172, 1991.
T. Poggio and F. Girosi, “Networks for approximation andlearning”, Proceedings of the IEEE, Vol. 78, No. 9, 1481–1497, 1990.
D.T.M. Slock, “On the convergence behavior of the LMS and the normalized LMSalgorithms”, IEEE Transactions on Signal Processing, Vol. 41, No. 9, pp. 2811–2825, Sept. 1993.
R.L. Streit and T.E. Luginbuhl, “Maximum likelihood training ofprobabilistic neural networks”, IEEE Transactions on Neural Networks, Vol. 5, No. 5, 764–783, 1994.
M.A. Styblinski and T.-S. Tang, “Experiments in nonconvexoptimization: stochastic approximation with function smoothing and simulated annealing”, Neural Networks, Vol. 3, 467–483, 1990.
J.A.K. Suykens, J.P.L. Vandewalle and B.L.R. De Moor, Artificial neural networks for modelling and control of nonlinear systems, Kluwer Academic Publishers: Boston, 1996.
J.A.K. Suykens and J. Vandewalle,“Nonconvex optimization using a Fokker-Planck Learning Machine”, 12th European Conference on Circuit Theory and Design (ECCTD 95), Istanbul Turkey, pp. 983-986, August 1995.
J.A.K. Suykens and J. Vandewalle, “A Fokker-Planck learning machine for globaloptimization”, submitted for publication.
N.G. van Kampen, Stochastic processes inphysics and chemistry, North-Holland, 1981.
F. Vanpoucke and M. Moonen,“Systolic robust adaptive beamforming with an adjustable constraint”, IEEE Transactions on Aerospace and Electronic Systems, Vol. 31, No. 2, pp. 658–669, April 1995.
B. Widrow and S.D. Stearns, Adaptive signal processing, PrenticeHall: EnglewoodCliffs, 1985.
E. Wong, Stochastic processes in information and dynamical systems,McGraw-Hill, 1971.
Author information
Authors and Affiliations
Rights and permissions
About this article
Cite this article
Suykens, J., Verrelst, H. & Vandewalle, J. On-Line Learning Fokker-Planck Machine. Neural Processing Letters 7, 81–89 (1998). https://doi.org/10.1023/A:1009632428145
Issue Date:
DOI: https://doi.org/10.1023/A:1009632428145