Pattern recognition based on simulated neural network.  (a) The SLP network with 784 input neurons and 10 output neurons. (b) 200 levels of synaptic weights were obtained from the optical-pulse and electric-pulse trains of our synaptic transistor. (c) The recognition rate of the network with the increasement of training epochs. (d) The hotspot graphs of synaptic weights to recognize the number “7” under various emotions.