Published-Work

This repository contains a list of all my published work

1. Wuraola, Adedamola, and Nitish Patel. "SQNL: A New Computationally Efficient Activation Function." 2018 International Joint Conference on Neural Networks (IJCNN). IEEE, 2018. Link: https://bit.ly/2SMT7Wp

2. Wuraola, Adedamola, and Nitish Patel. "Stochasticity-Assisted Training in Artificial Neural Network." International Conference on Neural Information Processing. Springer, Cham, 2018. Link: https://bit.ly/2TLksW0

3. Wuraola, Adedamola, and Nitish Patel. "Computationally Efficient Radial Basis Function." International Conference on Neural Information Processing. Springer, Cham, 2018. Link: https://bit.ly/2SMXikW

4. Wuraola, Adedamola, and Nitish Patel. "SQNL: A Universal Activation Function For Machine Learning." Under Review

LSTM

On LSTM architecture, I have tested the proposed function using the following datasets

1. Sentiment Analysis using the imdb dataset containining 60,000 reviews

2. Sequence to Sequence Translation using the English to French dataset saved under the filename fra.txt

3. Text Generation using the nietzsche text, found in 'https://s3.amazonaws.com/text-datasets/nietzsche.txt'

Autoencoders

Mnist Dataset using ELU vs SQLU