Title :
A winner-take-all circuit based on second order Hopfield neural networks as building blocks
Author :
Tymoshchuk, P. ; Kaszkurewicz, E.
Author_Institution :
Dept. of Electr. Eng., Fed. Univ. of Rio de Janeiro, Brazil
Abstract :
A new analog inhibitory Hopfield based winner-take-all (WTA) circuit is proposed. The general circuit structure is a binary tree arrangement of second order Hopfield networks and logic nodes. The connection matrix belongs to the class of diagonally stable block diagonal matrices and the activation functions are piecewise linear or sigmoidal. The mathematical justification of the WTA circuit functioning, on the basis of the uniqueness and global asymptotic stability of the equilibrium, is given. Simulation examples are also provided.
Keywords :
Hopfield neural nets; analogue processing circuits; asymptotic stability; circuit stability; matrix algebra; piecewise linear techniques; binary tree arrangement; connection matrix; diagonally stable block diagonal matrices; global asymptotic stability; logic nodes; mathematical justification; piecewise linear function; second order Hopfield neural networks; sigmoidal function; winner-take-all circuit; Asymptotic stability; Binary trees; Circuit simulation; Convergence; Hopfield neural networks; Logic circuits; Neural networks; Neurons; Piecewise linear techniques; Symmetric matrices;
Conference_Titel :
Neural Networks, 2003. Proceedings of the International Joint Conference on
Print_ISBN :
0-7803-7898-9
DOI :
10.1109/IJCNN.2003.1223808
