We aim to engineer reservoir computing using hardware chaotic neural networks, including associative memory recall. In this study, a reservoir layer was constructed using a pulse-type hardware chaotic neuron model (P-HCNM). The structure of the reservoir layer is simple and advantageous for hardware implementation. A hardware small-world neural network using a synaptic model and a gap junction model was constructed. The extent of memory recall in hardware reservoir computing was investigated using the reservoir layer. It was clarified that the P-HCNM could recall the image pattern when the recall signal was applied to the P-HCNM in an input layer, corresponding to the image pattern learned subsequently.
Yamaguchi et al. (Sat,) studied this question.