Applied Artificial Intelligence: Neural networks and deep learning with Python and TensorFlow

What are the secrets of modern Artificial Intelligence? How does AI beat humans in various domains, such as playing Go or predicting the future?
How can I implement my own Artificial Intelligence and push it into production with Google TensorFlow 2 and Keras? This book is about uncovering the basics of Artificial Neural Networks (ANN) and Deep Learning and how to implement AI models for production systems by using TensorFlow 2.
The first part of this book explains how to implement your own neural networks in Python from the scratch and to apply this technique to any given problem.
In step-by-step application examples, the reader learns how to implement neural networks in Python and to solve non-linear problems.
The book explains how neural networks are built, trained with sample data sets and how these networks are capable of solving complex problems.
The simplicity of the tutorial as well as the simple syntax of the Python language quickly enables the reader to transfer that knowledge and algorithms to any other programming language of choice.
The examples cover the design of simple neural networks for solving math functions or character recognition by using neural networks written in Python over solving simple regression problems to text analysis and classification.
The second part of the book shows how to build machine learning models in Google TensorFlow 2 and how to bring your Artificial Intelligence and machine-learning models into production.
TensorFlow is one of the most advanced open source machine learning frameworks available today. TensorFlow easily enables data scientists to push their Artificial Intelligence into a scalable production environment.
The third part of the book is dedicated to practical and fun machine learning examples, such as to calculate book recommendations or to predict the chance of survival for passengers of RMS Titanic.

Introduction
Artificial intelligence
Neural networks and deep learning
Activation of a neuron
Training a single neuron
Model a network of neurons
Use-case: Handwriting and character recognition
Taxonomy of machine learning use-cases
Classification
Regression
Clustering
AI in production with TensorFlow
System architecture
Distribution architecture
Run TensorFlow as a Docker container
Building your first computational graph
TensorFlow functions
Trace TensorFlow functions with TensorBoard
Implement a first linear regression model
Build a first model with Keras
Trace a Keras model using TensorBoard
Keras take-aways
Titanic: Can we train a model to predict survival?
Preparing and massaging the training data
Trace of the Titanic model training
Titanic survival prediction summary
Crowd Intelligence: Build your own book recommender
Load example book ratings
Building the Matrix Factorization Keras Model
Training the matrix factorization model
Final book to user rating and recommendation
Go beyond: The Book-Crossing dataset
Sentiment Analysis: How to detect Toxic Comments?
Retrieve and load the text data set
Representing text as numeric features
One-hot encoding
Word index encoding
Word embeddings
Training a word embedding
Predict toxic texts
Safe your trained word embedding to disk
Summary
References
Contact and download links
Credits