Abstract
Using short-time correlation function measurements of an observed process as input, we show that it is possible to train a network to learn the non-linear stochastic dynamics underlying the process. Alternatively this can be formulated as a neural network with non-linear stochastic synapses, which can, after training, be used to associate actions.
This is a preview of subscription content, log in via an institution to check access.
Preview
Unable to display preview. Download preview PDF.
Similar content being viewed by others
References
H.Haken, Information and Self-Organization, Springer-Verlag, Berlin-Heidelberg-New York 1988
L.Borland, H.Haken, Unbiased Determination of Forces Causing observed Processes, Z. Phys. B — Condensed Matter 81 (1992) 95
L.Borland, H. Haken, Unbiased Estimate of Forces from Measured Correlation Functions, including the Case of Strong Multiplicative Noise, Ann. Physik 1 (1992) 452
L.Borland, H. Haken, Learning the Dynamics of Two Dimensional Stochastic Markov Processes, to be published in Open Systems and Information Dynamics
E.T. Jaynes, Phs. Rev 106 (1957) 4,620
Author information
Authors and Affiliations
Editor information
Rights and permissions
Copyright information
© 1993 Springer-Verlag Berlin Heidelberg
About this paper
Cite this paper
Borland, L., Haken, H. (1993). Learning networks for process identification and associative action. In: Mira, J., Cabestany, J., Prieto, A. (eds) New Trends in Neural Computation. IWANN 1993. Lecture Notes in Computer Science, vol 686. Springer, Berlin, Heidelberg. https://doi.org/10.1007/3-540-56798-4_222
Download citation
DOI: https://doi.org/10.1007/3-540-56798-4_222
Published:
Publisher Name: Springer, Berlin, Heidelberg
Print ISBN: 978-3-540-56798-1
Online ISBN: 978-3-540-47741-9
eBook Packages: Springer Book Archive