×

Further results on controllability of recurrent neural networks. (English) Zbl 0914.93011

Summary: This paper studies controllability properties of recurrent neural networks. The new contributions are: (1) an extension of a previous result to a slightly different model, (2) a formulation and proof of a necessary and sufficient condition, and (3) an analysis of a low-dimensional case for which the hypotheses made in previous work do not apply.

MSC:

93B05 Controllability
92B20 Neural networks for/in biological studies, artificial life and related topics

References:

[1] Albertini, F.; Dai Pra, P., Forward accessibility for recurrent neural networks, IEEE Trans. Automat. Control, 40, 1962-1968 (1995) · Zbl 0856.93012
[2] Albertini, F.; Sontag, E. D., For neural networks, function determines form, Neural Networks, 6, 975-990 (1993)
[3] Y.-C. Chu, K. Glover, Gain-scheduling for systems with repeated scalar nonlinearities, IEEE Conf. Decision and Control, San Diego, December 1997.; Y.-C. Chu, K. Glover, Gain-scheduling for systems with repeated scalar nonlinearities, IEEE Conf. Decision and Control, San Diego, December 1997.
[4] Z. Feng, A.N. Michel, Uniqueness and stability of equilibria of a class of neural networks with applications to the Hopfield model, IEEE Conf. Decision and Control, San Diego, December 1997.; Z. Feng, A.N. Michel, Uniqueness and stability of equilibria of a class of neural networks with applications to the Hopfield model, IEEE Conf. Decision and Control, San Diego, December 1997.
[5] C.L. Giles, G.Z. Sun, H.H. Chen, Y.C. Lee, D. Chen, Higher order recurrent networks and grammatical inference, in: D.S. Touretzky (Ed.), Advances in Neural Information Processing Systems 2, Morgan Kaufmann, San Mateo, CA, 1990.; C.L. Giles, G.Z. Sun, H.H. Chen, Y.C. Lee, D. Chen, Higher order recurrent networks and grammatical inference, in: D.S. Touretzky (Ed.), Advances in Neural Information Processing Systems 2, Morgan Kaufmann, San Mateo, CA, 1990.
[6] K.A. Grasse, H.J. Sussmann, Global controllability by nice controls, in: H.J. Sussmann (Ed.), Nonlinear Controllability and Optimal Control, Dekker, New York, 1990, pp. 33-79.; K.A. Grasse, H.J. Sussmann, Global controllability by nice controls, in: H.J. Sussmann (Ed.), Nonlinear Controllability and Optimal Control, Dekker, New York, 1990, pp. 33-79. · Zbl 0703.93014
[7] K.J. Hunt, G.W. Irwin, K. Warwick (Eds.), Neural Network Engineering in Dynamic Control Systems, Springer, London, 1995.; K.J. Hunt, G.W. Irwin, K. Warwick (Eds.), Neural Network Engineering in Dynamic Control Systems, Springer, London, 1995.
[8] M. Krstic, I. Kanellakopoulos, P. Kokotovic, Nonlinear and Adaptive Control Design, Wiley, New York, 1995.; M. Krstic, I. Kanellakopoulos, P. Kokotovic, Nonlinear and Adaptive Control Design, Wiley, New York, 1995. · Zbl 0763.93043
[9] M. Matthews, A state-space approach to adaptive nonlinear filtering using recurrent neural networks, in: Proc. IASTED Symp. on Artificial Intelligence Applications and Neural Networks, Zürich, July 1990, pp. 197-200.; M. Matthews, A state-space approach to adaptive nonlinear filtering using recurrent neural networks, in: Proc. IASTED Symp. on Artificial Intelligence Applications and Neural Networks, Zürich, July 1990, pp. 197-200.
[10] Michel, A. N.; Farrell, J. A.; Porod, W., Qualitative analysis of neural networks, IEEE Trans. Circuits System, 36, 229-243 (1989) · Zbl 0672.94015
[11] M.M. Polycarpou, P.A. Ioannou, Neural networks and on-line approximators for adaptive control, in: Proc. 7th Workshop on Adaptive and Learning Systems, Yale University, 1992, pp. 93-798.; M.M. Polycarpou, P.A. Ioannou, Neural networks and on-line approximators for adaptive control, in: Proc. 7th Workshop on Adaptive and Learning Systems, Yale University, 1992, pp. 93-798.
[12] Sontag, E. D., Finite dimensional open-loop control generators for nonlinear systems, Int. J. Control, 47, 537-556 (1988) · Zbl 0641.93035
[13] E.D. Sontag, Mathematical Control Theory: Deterministic Finite Dimensional Systems, 2nd ed., Springer, New York, 1998.; E.D. Sontag, Mathematical Control Theory: Deterministic Finite Dimensional Systems, 2nd ed., Springer, New York, 1998. · Zbl 0945.93001
[14] E.D. Sontag, Neural networks for control, in: H.L. Trentelman, J.C. Willems (Eds.), Essays on Control: Perspectives in the Theory and its Applications, Birkhauser, Boston, 1993, pp. 339-380.; E.D. Sontag, Neural networks for control, in: H.L. Trentelman, J.C. Willems (Eds.), Essays on Control: Perspectives in the Theory and its Applications, Birkhauser, Boston, 1993, pp. 339-380. · Zbl 0844.93012
[15] E.D. Sontag, Recurrent neural networks: Some systems-theoretic aspects, in: M. Karny, K. Warwick, V. Kurkova (Eds.), Dealing with Complexity: a Neural Network Approach, Springer, London, 1997, pp. 1-12.; E.D. Sontag, Recurrent neural networks: Some systems-theoretic aspects, in: M. Karny, K. Warwick, V. Kurkova (Eds.), Dealing with Complexity: a Neural Network Approach, Springer, London, 1997, pp. 1-12.
[16] Sontag, E. D.; Sussmann, H. J., Complete controllability of continuous-time recurrent neural networks, Systems and Control Lett., 30, 177-183 (1997) · Zbl 0901.93005
[17] H.J. Sussmann, Local controllability and motion planning for some classes of systems with drift, in: Proc. 30th Conf. Decision and Control, Brighton, UK, IEEE Publications, New York, 1991, pp. 1110-1114.; H.J. Sussmann, Local controllability and motion planning for some classes of systems with drift, in: Proc. 30th Conf. Decision and Control, Brighton, UK, IEEE Publications, New York, 1991, pp. 1110-1114.
[18] R. Zbikowski, Lie algebra of recurrent neural networks and identifiability, in: Proc. Amer. Automat. Control Conf., San Francisco, 1993, pp. 2900-2901.; R. Zbikowski, Lie algebra of recurrent neural networks and identifiability, in: Proc. Amer. Automat. Control Conf., San Francisco, 1993, pp. 2900-2901.
This reference list is based on information provided by the publisher or from digital mathematics libraries. Its items are heuristically matched to zbMATH identifiers and may contain data conversion errors. In some cases that data have been complemented/enhanced by data from zbMATH Open. This attempts to reflect the references listed in the original paper as accurately as possible without claiming completeness or a perfect matching.