Dynamics of structured Complex Recurrent Hopfield Networks
Garimella Rama Murthy, Anil Rayala, Sai Dileep Munugoti · 2016
In this research paper, Complex Recurrent Hopfield Networks are introduced. Dynamics of a Structured Complex Recurrent Hopfield Network whose synaptic weight matrix is skew Hermitian is studied. It is proved that such a network when operated in parallel mode leads to a cycle of length 4. Two new types of complex matrices known as Braided Hermitian and Braided Skew Hermitian are defined and dynamics of a complex recurrent Hopfield network based on them is studied. It is reasoned that a cycle of length 8 is reached in parallel mode of operation. Experimental results for different types of Complex Recurrent Hopfield Networks are also discussed.