Prior knowledge input neural network method for GFET description

Abstract

For circuit design, various compact models for graphene field-effect transistors (GFETs) have been developed. However, the consistency between them is poor, since study on the mechanism of operation of GFETs is still immature and the models were derived based on different understandings. Herein, we propose another approach for circuit-level description of GFETs based on a prior knowledge input neural network modeling method. By virtue of the neural network's learning ability, it can accurately describe different GFETs without an exact description of their mechanism of operation. Basic knowledge on GFETs and an adaptive genetic algorithm are employed to improve the precision of the neural network, resulting in performance significantly exceeding that of a traditional, multilayer perceptron network. The universality of the method is verified by detailed tests using two different datasets. Its applicability for circuit design is demonstrated by relevant circuit simulations in a Verilog-A implementation.

Input significance analysis: feature ranking through synaptic weights manipulation for ANNS-based classifiers

“…It has been claimed as the best function generating weights and biases that increase the speed of the training [44]. It is also claimed that a better starting point to the algorithm used can be created by using the generated weights and biases for input to hidden layer nodes, obtained from this function [45].…”

Acceleration of nonequilibrium Green’s function simulation for nanoscale FETs by applying convolutional neural network model

“…Therefore it is important to reduce the computational time required for NEGF device simulations; for instance, by applying information-scientific viewpoint [11,12,13,14]. As a representative example, consider double-gate MOSFET (DG-MOSFET) [15,16,17,18,19,20,21].…”