Design and electrical simulation of on-chip neural learning based on nanocomponents

摘要

A neural inspired lookup table for reconfigurable circuits is described and simulated. The design is based on conductive bridge RAM to implement the synapses and carbon nanotube field effect transistors (CNTFET) for the other parts. Electrical simulations demonstrate compatibility between the nanocomponents and show the successful training of a linearly separable logical function NOR3.