Neural Networks: Neural Networks Tools and Techniques for Beginners

Neural Networks

Neural Networks Tools and Techniques for Beginners

Table of Contents

What are Artificial Neural Networks?

Fundamentals of Neural Networks

Activation Paradigms

Learning Paradigms

Multilayer Perceptron

Practical Application – Text Recognition

Practical Application – Image Processing

Problems with Neural Networks

Conclusion

DISCLAIMER

Copyright © 2018 John Slavio

All Rights Reserved

No part of this eBook can be transmitted or reproduced in any form including print, electronic, photocopying, scanning, mechanical, or recording without prior written permission from the author.

While the author has taken the utmost effort to ensure the accuracy of the written content, all readers are advised to follow information mentioned herein at their own risk. The author cannot be held responsible for any personal or commercial damage caused by information. All readers are encouraged to seek professional advice when needed.

ABOUT THE AUTHOR

John Slavio is a programmer who is passionate about the reach of the internet and the interaction of the internet with daily devices. He has automated several home devices to make them 'smart' and connect them to high speed internet. His passions involve computer security, iOT, hardware programming and blogging. 

What are Artificial Neural Networks?

Artificial Neural Networks are digitized nodes designed to mimic the biological nodes of the brain to complete computerized tasks. That’s about as tiny of a definition for ANN’s as you will ever see, but normally when one asks the what question, they also ask how does it work because that usually provides a better explanation of the what question. Therefore, the rest of this book will cover the how does for neural networks.

We use Python Here and Here’s Why

It’s Easier to Conceptualize In

The reason why Java and Python are really popular within college environments is due to the fact that they are easier for conceptualizing various computer science topics. The reason why we use Python within this book to explore code examples is due to the fact that it has a relatively simplistic concept that can be expressed through very meaningful syntactical words.

It’s the Most Popular Option

Python is also the most popular option for handling any advanced machine learning or computer vision algorithm. This is particularly because Python is a primary language used by science researchers, which means that along with the C language, Python receives special attention from the inventors of these algorithms.

It has the Most Access

As it is the most popular, it will also have the most access when it comes to different algorithms. While algorithms like the A * algorithm is a popular algorithm amongst many languages, subjects like Hough Lines or Hough Contours are rather uncommon in languages like JavaScript. Even though JavaScript is a fantastic language, it isn't one of those languages that are on the forefront for handling computer vision and neural networking. Sure, there is a JavaScript solution for these things and there is a tensor.js that came out not too long ago, but Python is the one that's going to be capable of handling a lot more functions.

Let’s start with Algorithms

A Deceiving Algorithm: The A * Algorithm

This may look like a foreign concept to you, but if you have played any video game that involves an enemy attempting to follow you, you have encountered this algorithm. Specifically, imagine you have soldiers and they need to get to you as quickly as possible in order to keep you engaged in whatever game you are playing. That's right, this algorithm represents what is known as pathfinding. n is the last node on the path, g(n) is the cost of the path from the start to n, and h(n) is the estimated cost to the end of the path (using the shortest way).

Pathfinding is a really deceiving algorithm that looks like it could be machine learning. The problem is that pathfinding is a Brute Force tactic with some clever recursive algebraic equations that requires the algorithm to know where the end of the algorithm is. Therefore, once your enemy soldier spawns into the room, it begins with two basic variables. You have the variable that contains where the enemy soldier is in the room and you have where you are in the room. Normally, these are considered X & Y coordinates.

What the algorithm does is it finds the shortest path to you but because it is clearly evident when an enemy simply runs to you, developers employ deceiving tactics. You see, essentially, they allow the pathfinding variable to take breaks or get within a certain radius so that the enemy can stop and fire behind a cover of some sort. This makes it seem like an enemy soldier has