Title :
Synthesis for symmetric weight matrices of neural networks
Author :
Saubhayana, M. ; Newcomb, R.W.
Author_Institution :
Dept. of Electr. & Comput. Eng., Maryland Univ., College Park, MD, USA
Abstract :
A synthesis method to guarantee symmetric weight matrices for a class of neural networks (which includes the Hopfield neural network as a special case) is proposed. This fills in a gap in the Li-Michel-Porod´s synthesis and guarantees asymptotic stability for a given set of linearly independent equilibrium points under Lyapunov´s stability criteria.
Keywords :
Hopfield neural nets; asymptotic stability; stability criteria; Hopfield neural network; Li-Michel-Porod´s synthesis; asymptotic stability; linearly independent equilibrium points; stability criteria; symmetric weight matrices; Asymptotic stability; Educational institutions; Hopfield neural networks; Matrix decomposition; Network synthesis; Neural networks; Nonlinear equations; Stability analysis; Symmetric matrices; Voltage;
Conference_Titel :
Circuits and Systems, 2003. ISCAS '03. Proceedings of the 2003 International Symposium on
Print_ISBN :
0-7803-7761-3
DOI :
10.1109/ISCAS.2003.1206403
