Deep Learning with Keras: Beginner’s Guide to Deep Learning with Keras

By Frank Millstein

This book will introduce you to various supervised and unsupervised deep learning algorithms like the multilayer perceptron, linear regression and other more advanced deep convolutional and recurrent neural networks. You will also learn about image processing, handwritten recognition, object recognition and much more.

Furthermore, you will get familiar with recurrent neural networks like LSTM and GAN as you explore processing sequence data like time series, text, and audio.

The book will definitely be your best companion on this great deep learning journey with Keras introducing you to the basics you need to know in order to take next steps and learn more advanced deep neural networks.

Here Is a Preview of What You'll Learn Here…

• The difference between deep learning and machine learning
• Deep neural networks
• Convolutional neural networks
• Building deep learning models with Keras
• Multi-layer perceptron network models
• Activation functions
• Handwritten recognition using MNIST
• Solving multi-class classification problems
• Recurrent neural networks and sequence classification
• And much more...

Get this book NOW and learn more about Deep Learning with Keras!

• Language: English
• Publisher: Frank Millstein
• Release date: Oct 20, 2019
• ISBN: 9781393142942

Book preview

By Frank Millstein

WHAT IS IN THE BOOK?

INTRODUCTION

HOW DEEP LEARNING IS DIFFERENT FROM MACHINE LEARNING

DEEPER INTO DEEP LEARNING

CHAPTER 1: A FIRST LOOK AT NEURAL NETWORKS

CONVOLUTIONAL NEURAL NETWORK

RECURRENT NEURAL NETWORK

RNN SEQUENCE TO SEQUENCE MODEL

AUTOENCODERS

REINFORCEMENT DEEP LEARNING

GENERATIVE ADVERSARIAL NETWORK

CHAPTER 2: GETTING STARTED WITH KERAS

BUILDING DEEP LEARNING MODELS WITH KERAS

CHAPTER 3: MULTI-LAYER PERCEPTRON NETWORK MODELS

MODEL LAYERS

MODEL COMPILATION

MODEL TRAINING

MODEL PREDICTION

CHAPTER 4: ACTIVATION FUNCTIONS FOR NEURAL NETWORKS

SIGMOID ACTIVATION FUNCTION

TANH ACTIVATION FUNCTION

RELU ACTIVATION FUNCTION

CHAPTER 5: MNIST HANDWRITTEN RECOGNITION

CHAPTER 6: NEURAL NETWORK MODELS FOR MULTI-CLASS CLASSIFICATION PROBLEMS

ONE-HOT ENCODING

DEFINING NEURAL NETWORK MODELS WITH SCIKIT-LEARN

EVALUATING MODELS WITH K-FOLD CROSS VALIDATION

CHAPTER 7: RECURRENT NEURAL NETWORKS

SEQUENCE CLASSIFICATION WITH LSTM RECURRENT NEURAL NETWORKS

WORD EMBEDDING

APPLYING DROPOUT

NATURAL LANGUAGE PROCESSING WITH RECURRENT NEURAL NETWORKS

LAST WORDS

Copyright © 2018 by Frank Millstein- All rights reserved.

This document is geared towards providing exact and reliable information in regards to the topic and issue covered. The publication is sold with the idea that the publisher is not required to render accounting, officially permitted, or otherwise, qualified services. If advice is necessary, legal or professional, a practiced individual in the profession should be ordered.

From a Declaration of Principles which was accepted and approved equally by a Committee of the American Bar Association and a Committee of Publishers and Associations.

In no way is it legal to reproduce, duplicate, or transmit any part of this document by either electronic means or in printed format. Recording of this publication is strictly prohibited, and any storage of this document is not allowed unless with written permission from the publisher. All rights reserved.

The information provided herein is stated to be truthful and consistent, in that any liability, in terms of inattention or otherwise, by any usage or abuse of any policies, processes, or directions contained within is the solitary and utter responsibility of the recipient reader. Under no circumstances will any legal responsibility or blame be held against the publisher for any reparation, damages, or monetary loss due to the information herein, either directly or indirectly.

Respective authors own all copyrights not held by the publisher.

The information herein is offered for informational purposes solely and is universal as so. The presentation of the information is without contract or any type of guarantee assurance.

The trademarks that are used are without any consent, and the publication of the trademark is without permission or backing by the trademark owner. All trademarks and brands within this book are for clarifying purposes only and are owned by the owners themselves, not affiliated with this document.

INTRODUCTION

Neural networks and deep learning are increasingly important studies and concepts in computer science with amazing strides being made by major tech companies like Google. Over the years, you may have heard words like backpropagation, neural networks, and deep learning tossed around a lot. Therefore, as we hear them more often, there is little wonder why these terms have seized your curiosity.

Deep learning is an important area of active research today in the field of computer science. If you are involved in this scientific area, I am sure you have come across these terms at least once. Deep learning and neural networks may be an intimidating concept, but since it is increasingly popular these days, this topic is most definitely worth your attention.

Google and other large global tech companies are making great strides with deep-learning projects, like the Google Brain project and its recent acquisition called DeepMind. Moreover, many deep learning methods are beating those traditional machine learning methods on every single matric.

HOW DEEP LEARNING IS DIFFERENT FROM MACHINE LEARNING

Before going further into this subject, we must take a step back, so you get to learn more about the broader field of machine learning. Very often, we encounter problems for which it is hard to write a computer program for solving those issues. For instance, if you want to program your computer to recognize specific handwritten digits that you may encounter on certain issues, you can try to devise a collection of rules to distinguish every individual digit. In this case, zeros are one closed loop, but what if you did not perfectly close this loop. On the other hand, what if the right top of your loop closes on that part where the left top of your loop starts?

Issues like this happen routinely, as zero may be very difficult when it comes to distinguishing from six algorithmically. Therefore, you have issues when differentiating zeroes from sixes. You could establish a kind of cutoff, but you will have problems deciding the origination of the cutoff in the first place. Therefore, quickly it becomes very complicated to compile a list of guesses and rules that will accurately classify your handwritten digits.

There are many more kinds of issues that fall into this category such as comprehending speech, recognizing objects, and understanding concepts. Therefore, we can have issues when writing computer programs, as we do not know how this is done by human brains. Despite the fact you have a relatively good idea on how to do this, your program may be very complicated.

Therefore, instead of writing a program, you can try