awur978/Recurrent-Neural-Network

Stars
2
Language
Jupyter Notebook
Created about 6 years ago
Updated almost 5 years ago

awur978/Recurrent-Neural-Network

awur978

There are no reviews yet. Be the first to send feedback to the community and the maintainers!

This repo contains the LSTM architecture trained on proposed Activation function

Levenberg_Marquardt-toolbox

This is a matlab code for LM algorithm use for training Artificial Neural Network

Time_Series_Application

This is a repo containing research work on NN speed control of DC motor using MATLAB. The aim is to implement this on FPGA. Overall, we are interested in comparing different architectures, delays, sampling time and finally activation functions.

Activation-Functions

This is a repository containing the new activation function proposed during the cause of my PhD

Published-Work

All published Work

Jupyter Notebook

SQ_RBF_Paper

msu

Effect of parameterizing square-based asymmetric activation functions

Stochasticity_Paper

This repo contains the code we used for the paper titled "Stochasticity-Assisted Training in Artificial Neural Network" presented at The 25th International Conference on Neural Information Processing (ICONIP 2018)

Autoencoder

This repo contains the Keras code for fully connected autoencoder found in several articles.

Stochasticity_LM

Effect of introducing stochasticity in NN convergence and helping with neuron saturation

Hardware_Activation-Functions

This contains all the VHDL code used for our novel Activation Functions

Approximation_AF

This repository contains all the different approximations method in literature for hardware implementation of bipolar nonlinear function. Focus is on Tanh activation function. Codes are initially written in python, MATLAB and then the VHDL description language is use for hardware implementation of this approximation methods