Two-Step Implementation of Sigmoid Function for Artificial Neural Network in Field Programmable Gate Array
In this paper, a combination of second order nonlinear function (SONF) and differential look-up table (differential LUT) is introduced as a sigmoid function for implementing the artificial neural network (ANN) in field programmable gate array (FPGA). Implementing ANN on FPGA will overcome the slow r...
Main Authors: | , , |
---|---|
Format: | Conference or Workshop Item |
Language: | English |
Published: |
2014
|
Subjects: | |
Online Access: | http://umpir.ump.edu.my/id/eprint/6903/ http://umpir.ump.edu.my/id/eprint/6903/1/Two-Step_Implementation_of_Sigmoid_Function_for_Artificial_Neural_Network_in_Field_Programmable_Gate_Array.pdf |
Summary: | In this paper, a combination of second order nonlinear function (SONF) and differential look-up table (differential LUT) is introduced as a sigmoid function for implementing the artificial neural network (ANN) in field programmable gate array (FPGA). Implementing ANN on FPGA will overcome the slow response for real-time application and portable issues that arise in the software-based ANN. The output accuracy achieved by this two-step approach is ten times better than that of using only SONF and two times better than that of using conventional LUT. Thus the proposed idea is suitable to be implemented as a hardware-based ANN for various real-time applications. |
---|