Hands-On Neural Networks: Learn how to build and train your first neural network model using Python By Leonardo De Marchi, Laura Mitchell ¨PDF

Hands-On Neural Networks: Learn how to build and train your first neural network model using Python By Leonardo De Marchi, Laura Mitchell ¨PDF

Informations about the book:

Title: Hands-On Neural Networks: Learn how to build and train your first neural network model using Python

Author: Lentin Joseph

Size: 22.8

Format: EPUB

Year: 2019

Pages: 269

Book Contents:

Section 1: Getting Started

Chapter 1: Getting Started with Supervised Learning 

History of AI 

An overview of machine learning

Supervised learning 

Unsupervised learning 

Semi-supervised learning 

Reinforcement learning 

Environment setup 

Understanding virtual environments 

Anaconda 

Docker 

Supervised learning in practice with Python 

Data cleaning 

Feature engineering 

How deep learning performs feature engineering 

Feature scaling 

Feature engineering in Keras 

Supervised learning algorithms 

Metrics 

Regression metrics 

Classification metrics 

Evaluating the model 

TensorBoard 

Summary 

Chapter 2: Neural Network Fundamentals 

The perceptron 

Implementing a perceptron 

Keras 

Implementing perceptron in Keras 

Feedforward neural networks 

Introducing backpropagation 

Activation functions 

Sigmoid 

Softmax 

Tanh 

ReLU 

Table of Contents

[ ii ]

Keras implementation 

The chain rule 

The XOR problem 

FFNN in Python from scratch 

FFNN Keras implementation 

TensorBoard 

TensorBoard on the XOR problem 

Summary 

Section 2: Deep Learning Applications

Chapter 3: Convolutional Neural Networks for Image Processing 

Understanding CNNs 

Input data 

Convolutional layers 

Pooling layers 

Stride 

Max pooling 

Zero padding 

Dropout layers 

Normalization layers 

Output layers 

CNNs in Keras 

Loading the data 

Creating the model 

Network configuration 

Keras for expression recognition 

Optimizing the network 

Summary 

Chapter 4: Exploiting Text Embedding 

Machine learning for NLP 

Rule-based methods 

Understanding word embeddings 

Applications of words embeddings 

Word2vec 

Word embedding in Keras 

Pre-trained network 

GloVe 

Global matrix factorization 

Using the GloVe model 

Text classification with GloVe 

Summary 

Chapter 5: Working with RNNs 

Understanding RNNs 

Theory behind CNNs 

Table of Contents

[ iii ]

Types of RNNs 

One-to-one 

One-to-many 

Many-to-many 

The same lag 

A different lag 

Loss functions 

Long Short-Term Memory 

LSTM architecture 

LSTMs in Keras 

PyTorch basics 

Time series prediction 

Summary 

Chapter 6: Reusing Neural Networks with Transfer Learning 

Transfer learning theory 

Introducing multi-task learning

Reusing other networks as feature extractors 

Implementing MTL 

Feature extraction 

Implementing TL in PyTorch 

Summary 

Section 3: Advanced Applications

Chapter 7: Working with Generative Algorithms 

Discriminative versus generative algorithms 

Understanding GANs 

Training GANs 

GAN challenges 

GAN variations and timelines 

Conditional GANs 

DCGAN 

ReLU versus Leaky ReLU 

DCGAN – a coded example 

Pix2Pix GAN 

StackGAN 

CycleGAN 

ProGAN 

StarGAN 

StarGAN discriminator objectives 

StarGAN generator functions 

BigGAN 

StyleGAN 

Style modules 

StyleGAN implementation 

Deepfakes 

Table of Contents

[ iv ]

RadialGAN 

Summary 

Further reading 

Chapter 8: Implementing Autoencoders 

Overview of autoencoders 

Autoencoder applications 

Bottleneck and loss functions 

Standard types of autoencoder

Undercomplete autoencoders 

Example 

Visualizing with TensorBoard 

Visualizing reconstructed images 

Multilayer autoencoders 

Example 

Convolutional autoencoders 

Example 

Sparse autoencoders 

Example 

Denoising autoencoders 

Example 

Contractive autoencoder 

Variational Autoencoders 

Training VAEs 

Example 

Summary 

Further reading 

Chapter 9: Deep Belief Networks 

Overview of DBNs 

BBNs 

Predictive propagation 

Retrospective propagation 

RBMs 

RBM training 

Example – RBM recommender system 

Example – RBM recommender system using code 

DBN architecture 

Training DBNs 

Fine-tuning 

Datasets and libraries 

Example – supervised DBN classification 

Example – supervised DBN regression 

Example – unsupervised DBN classification 

Summary 

Further reading 

Table of Contents

[ v ]

Chapter 10: Reinforcement Learning 

Basic definitions 

Introducing Q-learning 

Learning objectives 

Policy optimization 

Methods of Q-learning 

Playing with OpenAI Gym 

The frozen lake problem 

Summary 

Chapter 11: Whats Next? 

Summarizing the book 

Future of machine learning 

Artificial general intelligence 

Ethics in AI 

Interpretability 

Automation 

AI safety 

AI ethics 

Accountability 

Conclusions