A Spiking Neural Network implemented with Single-Electron Transistors and NoCs

Publication date: September 2018Source: Nano Communication Networks, Volume 17Author(s): Beatriz dos Santos Pês, Janaina Gonçalves Guimarães, Elder Oroski, Marlio José do Couto BonfimAbstractThis work proposes a Spiking Neural Network, SNN, based on a nanoelectronic spiking neuron – as building block – and a 2D-mesh network-on-chip, NoC — as interconnect architecture. The SNN obtained from the NoC is an alternative for high density architectures, providing reconfigurability, high scalability and low power consumption. A look-up table router was used to connect all units. Further on, the eXclusive-OR, XOR, benchmark problem was used to validate the functionality of the nanoelectronic SNN.
Source: Nano Communication Networks - Category: Nanotechnology Source Type: research