Hardware implementation of the neural computation based on stochastic logic using single flux quantum circuits

点击翻译按钮获取中文摘要

Circuits for a hardware neural computation were fabricated by using single flux-quantum circuits based on stochastic logic. In the stochastic logic, the data is represented by a random pulse generation rate for a certain period on time domain. Main components of this neuron circuit are a multiplication of synaptic weight, a generation of membrane potential, and an activation function. These basic elements of the neuron circuit were fabricated. Some main circuits were successfully demonstrated.