Abstract
In this paper, two methods named Backward Computation (BC) and Forward Computation (FC) for both on-line and batch backward gradient computation of a system output (for sensitivity analysis) or cost function (for learning) with respect to system parameters are derived by the Signal-Flow-Graph representation theory and its known properties. The system can be any causal, in general non-linear and time-variant, dynamic system represented by a SFG, in particular any feedforward, time delay or recurrent neural network In this work, we use discrete time notation, but the same theory holds for the continuous time case. The gradient is obtained in a straightforward way by the analysis of two SFGs, the original one and its adjoint (for the BC method) or its derivative (FC method) both obtained from the first by simple transformations without the complex chain rule expansions of derivatives usually employed.
The BC and FC methods are dual and the adjoint and derivative SFGs (of the same SFG) can be obtained one from the other by graph transformations. The BC method is local in space but not in time while the FC is local in time but not in space.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
Preview
Unable to display preview. Download preview PDF.
References
Almeida, L. B., 1987, “A learning rule for asynchronous perceptrons with feedback in combinatorial environment”, Proc. Int. Conf. on Neural Networks, vol. 2, pp. 609–618.
Back A.D., Tsoi A.C., 1991, “FIR and IIR synapses, a new neural network architecture for time series modelling”, Neural Computation 3: 375–385.
Back A.D., Tsoi A.C., 1993, “A simplified gradient algorithm for IIR synapse Multilayer Perceptron”, Neural Computation 5: 456–462.
Beaufays F., Wan E., 1994, “Relating Real-Time Backpropagation and Backpropagation-Through-Time: An Application of Flow Graph Interreciprocity”, Neural Computation 6: 296–306.
Campolucci P., 1998, “A circuit theory approach to recurrent neural networks architectures and learning methods”, PhD dissertation, University of Bologna, Italy, 1998.
Campolucci P., Marchegiani A., Uncini A., Piazza F., 1997, “Signal-Flow-Graph derivation of on-line gradient learning algorithms” Proc. of ICNN-97, IEEE Int. Conference on Neural Networks, Houston (USA), June 1997.
Campolucci P., Piazza F., Uncini A., 1995, “On-line learning algorithms for neural networks with IIR synapses”, Proc. of the IEEE International Conference of Neural Networks, Perth, November 1995.
Campolucci P., Uncini A., Piazza F., 1998, “Dynamical Systems Learning by a Circuit Theoretic Approach”, Proc. of ISCAS-98, IEEE Int. Symposium on Circuits and Systems, 1998.
Campolucci P., Uncini A., Piazza F., Rao B.D., 1999, “On-line learning algorithms for locally recurrent neural networks”, IEEE Transactions on Neural Networks, Vol. 10, No. 2, March 1999.
Director S.W., Rohrer R.A., 1969, “The generalized adjoint network and network sensitivities”, IEEE Trans. on Circuit Theory, vol. CT-16, pp. 318–323, Aug. 1969.
Gherrity M., 1989, “A learning algorithm for analog, fully recurrent neural networks”, Proc. Int. Joint Conference Neural Networks, I, 643–644.
Griewank A., 1988, “On Automatic Differentiation”, Mathematics and Computer Science Division, Preprint ANL/MCS-P10-1088, November 1988.
Haykin S., 1994, “Neural Networks: a comprehensive foundation”, IEEE Press-Macmillan.
Horne B.G. and Giles C.L., “An experimental comparison of recurrent neural networks”, in Advances in Neural Information Processing Systems vol. 7, Cambridge, MA: MIT Press 1995 p. 697.
Jabri M. and Flower B., 1992, “Weight Perturbation: An Optimal Architecture and Learning Technique for Analog VLSI Feedforward and Recurrent Multilayer Networks”, IEEE Trans. on Neural Networks, vol. 3, no. 1, pp. 154–157, January 1992.
Lee A.Y., 1974, “Signal Flow Graphs—Computer-Aided System Analysis and Sensitivity Calculations”, IEEE Transactions on Circuits and Systems, vol. cas-21, no. 2, pp. 209–216, March 1974.
Martinelli, G. and Perfetti, R., 1991, “Circuit theoretic approach to the backpropagation learning algorithm”, Proc. of the IEEE Int. Symposium on Circuits and Systems, 1991.
Mason S.J., 1953, “Feedback theory-Some properties of signal-flow graphs”, Proc. of the Institute of Radio Engineers vol. 41, pp. 1144–1156.
Mason S.J., 1956, “Feedback theory-Further properties of signal-flow graphs”, Proc. of the Institute of Radio Engineers vol. 44, pp. 920–926.
Narendra K.S., Parthasarathy K., 1991, “Gradient Methods for the Optimization of Dynamical Systems Containing Neural Networks”, IEEE Trans. on Neural Networks, vol. 2, no. 2, March 1991.
Nerrand O., Roussel-Ragot P., Personnaz L., Dreyfus G., Marcos S., 1993, “Neural Networks and Nonlinear Adaptive Filtering: Unifying Concepts and New Algorithms”, Neural Computation 5: 165–199.
Oppenheim, A. V. and Schafer, R. W., 1975, “Digital Signal Processing”, Prentice-Hall.
Osowski, S. 1994, “Signal flow graphs and neural networks”, Biological Cybernetics, 70, pp. 387–395.
Pearlmutter, B. A., 1995, “Gradient Calculations for Dynamic Recurrent Neural Networks: a Survey”, IEEE Trans. on Neural Networks, vol. 6, no. 5.
Penfield, P., Spence, R. and Duiker, S., 1970, “Tellegen’s Theorem and Electrical Networks”, MIT Press, Cambridge, MA.
Shynk J.J., 1989, “Adaptive IIR filtering”, IEEE ASSP Magazine, April 1989.
Srinivasan B., Prasad U.R. and Rao N.J., 1994, “Backpropagation through Adjoints for the identification of Non linear Dynamic Systems using Recurrent Neural Models”, IEEE Trans. on Neural Networks pp. 213–228, March 1994.
Tellegen, B.D.H., 1952, “A general network theorem, with applications”, Philips Res. Rep. 7, 259–269.
Tsoi A.C., Back A.D., 1994, “Locally recurrent globally feedforward networks: a critical review of architectures”, IEEE Transactions on Neural Networks, vol. 5, no. 2, 229–239, March 1994.
Uncini A., Vecci L., Campolucci P., Piazza F., 1999, “Complex-valued Neural Networks with adaptive spline activation function for digital radio links nonlinear equalization”, IEEE Transactions on Signal Processing, Vol. 47, No. 2, February 1999.
Wan E.A., Beaufays F., 1996, “Diagrammatic Derivation of Gradient Algorithms for Neural Networks”, Neural Computation 8: 182–201.
Wan E.A., Beaufays F., 1998, “Diagrammatic methods for deriving and relating temporal neural networks algorithms” In adaptive Processing of Sequences and Data Structures, Lecture Notes in Artificial Intelligence — Springer Verlag 1998.
Werbos P.J., 1974, “Beyond regression: New tools for prediction and analysis in the behavioural sciences”, Ph.D. dissertation, Committee on Appi. Math., Harvard Univ., Cambridge, MA, Nov. 1974.
Werbos, P. J., 1990, “Backpropagation through time: what it does and how to do it”, Proc. of IEEE, Special issue on neural networks, vol. 78, no. 10, pp. 1550–1560.
Williams R.J., Peng J., 1990, “An efficient gradient-based algorithm for on line training of recurrent network trajectories”, Neural Computation 2: 490–501.
Williams R.J., Zipser D., 1989, “A Learning Algorithm for Continually Running Fully Recurrent Neural Networks”, Neural Computation 1: 270–280.
Williams, R. J. and Zipser, D., 1994, “Gradient-based learning algorithms for recurrent networks and their computational complexity”, in Backpropagation: Theory, Architectures and Applications, Y. Chauvin and D.E. Rumelhart, Eds. Hillsdale, NJ: Lawrence Erlbaum Associates, pp. 433–486.
Author information
Authors and Affiliations
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 1999 Springer-Verlag London Limited
About this paper
Cite this paper
Campolucci, P. (1999). Sensitivity Analysis and Learning of Non-Linear Dynamic Systems by Two Dual Signal-Flow-Graph Approaches. In: Marinaro, M., Tagliaferri, R. (eds) Neural Nets WIRN Vietri-99. Perspectives in Neural Computing. Springer, London. https://doi.org/10.1007/978-1-4471-0877-1_5
Download citation
DOI: https://doi.org/10.1007/978-1-4471-0877-1_5
Publisher Name: Springer, London
Print ISBN: 978-1-4471-1226-6
Online ISBN: 978-1-4471-0877-1
eBook Packages: Springer Book Archive