Deep Learning with R for Beginners电子书

About Packt

Why subscribe?

Packt.com

Contributors

About the authors

Packt is searching for authors like you

Preface

Who this book is for

What this book covers

To get the most out of this book

Download the example code files

Download the color images of this book

Conventions used

Get in touch

Reviews

Getting Started with Deep Learning

What is deep learning?

A conceptual overview of neural networks

Neural networks as an extension of linear regression

Neural networks as a network of memory cells

Deep neural networks

Some common myths about deep learning

Setting up your R environment

Deep learning frameworks for R

MXNet

Keras

Do I need a GPU (and what is it, anyway)?

Setting up reproducible results

Summary

Training a Prediction Model

Neural networks in R

Building neural network models

Generating predictions from a neural network

The problem of overfitting data – the consequences explained

Use case – building and applying a neural network

Summary

Deep Learning Fundamentals

Building neural networks from scratch in R

Neural network web application

Neural network code

Back to deep learning

The symbol, X, y, and ctx parameters

The num.round and begin.round parameters

The optimizer parameter

The initializer parameter

The eval.metric and eval.data parameters

The epoch.end.callback parameter

The array.batch.size parameter

Using regularization to overcome overfitting

L1 penalty

L1 penalty in action

L2 penalty

L2 penalty in action

Weight decay (L2 penalty in neural networks)

Ensembles and model-averaging

Use case – improving out-of-sample model performance using dropout

Summary

Training Deep Prediction Models

Getting started with deep feedforward neural networks

Activation functions

Introduction to the MXNet deep learning library

Deep learning layers

Building a deep learning model

Use case – using MXNet for classification and regression

Data download and exploration

Preparing the data for our models

The binary classification model

The regression model

Improving the binary classification model

The unreasonable effectiveness of data

Summary

Image Classification Using Convolutional Neural Networks

CNNs

Convolutional layers

Pooling layers

Dropout

Flatten layers, dense layers, and softmax

Image classification using the MXNet library

Base model (no convolutional layers)

LeNet

Classification using the fashion MNIST dataset

References/further reading

Summary

Tuning and Optimizing Models

Evaluation metrics and evaluating performance

Types of evaluation metric

Evaluating performance

Data preparation

Different data distributions

Data partition between training, test, and validation sets

Standardization

Data leakage

Data augmentation

Using data augmentation to increase the training data

Test time augmentation

Using data augmentation in deep learning libraries

Tuning hyperparameters

Grid search

Random search

Use case—using LIME for interpretability

Model interpretability with LIME

Summary

Natural Language Processing Using Deep Learning

Document classification

The Reuters dataset

Traditional text classification

Deep learning text classification

Word vectors

Comparing traditional text classification and deep learning

Advanced deep learning text classification

1D convolutional neural network model

Recurrent neural network model

Long short term memory model

Gated Recurrent Units model

Bidirectional LSTM model

Stacked bidirectional model

Bidirectional with 1D convolutional neural network model

Comparing the deep learning NLP architectures

Summary

Deep Learning Models Using TensorFlow in R

Introduction to the TensorFlow library

Using TensorBoard to visualize deep learning networks

TensorFlow models

Linear regression using TensorFlow

Convolutional neural networks using TensorFlow

TensorFlow estimators and TensorFlow runs packages

TensorFlow estimators

TensorFlow runs package

Summary

Anomaly Detection and Recommendation Systems

What is unsupervised learning?

How do auto-encoders work?

Regularized auto-encoders

Penalized auto-encoders

Denoising auto-encoders

Training an auto-encoder in R

Accessing the features of the auto-encoder model

Using auto-encoders for anomaly detection

Use case – collaborative filtering

Preparing the data

Building a collaborative filtering model

Building a deep learning collaborative filtering model

Applying the deep learning model to a business problem

Summary

Running Deep Learning Models in the Cloud

Setting up a local computer for deep learning

How do I know if my model is training on a GPU?

Using AWS for deep learning

A brief introduction to AWS

Creating a deep learning GPU instance in AWS

Creating a deep learning AMI in AWS

Using Azure for deep learning

Using Google Cloud for deep learning

Using Paperspace for deep learning

Summary

The Next Level in Deep Learning

Image classification models

Building a complete image classification solution

Creating the image data

Building the deep learning model

Using the saved deep learning model

The ImageNet dataset

Loading an existing model

Transfer learning

Deploying TensorFlow models

Other deep learning topics

Generative adversarial networks

Reinforcement learning

Summary

Handwritten Digit Recognition using Convolutional Neural Networks

What is deep learning and why do we need it?

What makes deep learning special?

What are the applications of deep learning?

Handwritten digit recognition using CNNs

Get started with exploring MNIST

First attempt – logistic regression

Going from logistic regression to single-layer neural networks

Adding more hidden layers to the networks

Extracting richer representation with CNNs

Summary

Traffic Signs Recognition for Intelligent Vehicles

How is deep learning applied in self-driving cars?

How does deep learning become a state-of-the-art solution?

Traffic sign recognition using CNN

Getting started with exploring GTSRB

First solution – convolutional neural networks using MXNet

Trying something new – CNNs using Keras with TensorFlow

Reducing overfitting with dropout

Dealing with a small training set – data augmentation

Reviewing methods to prevent overfitting in CNNs

Summary

Fraud Detection with Autoencoders

Getting ready

Installing Keras and TensorFlow for R

Installing H2O

Our first examples

A simple 2D example

Autoencoders and MNIST

Outlier detection in MNIST

Credit card fraud detection with autoencoders

Exploratory data analysis

The autoencoder approach – Keras

Fraud detection with H2O

Exercises

Variational Autoencoders

Image reconstruction using VAEs

Outlier detection in MNIST

Text fraud detection

From unstructured text data to a matrix

From text to matrix representation — the Enron dataset

Autoencoder on the matrix representation

Exercises

Summary

Text Generation using Recurrent Neural Networks

What is so exciting about recurrent neural networks?

But what is a recurrent neural network, really?

LSTM and GRU networks

LSTM

GRU

RNNs from scratch in R

Classes in R with R6

Perceptron as an R6 class

Logistic regression

Multi-layer perceptron

Implementing a RNN

Implementation as an R6 class

Implementation without R6

RNN without derivatives — the cross-entropy method

RNN using Keras

A simple benchmark implementation

Generating new text from old

Exercises

Summary

Sentiment Analysis with Word Embedding

Warm-up – data exploration

Working with tidy text

The more, the merrier – calculating n-grams instead of single words

Bag of words benchmark

Preparing the data

Implementing a benchmark – logistic regression

Exercises

Word embeddings

word2vec

GloVe

Sentiment analysis from movie reviews

Data preprocessing

From words to vectors

Sentiment extraction

The importance of data cleansing

Vector embeddings and neural networks

Bi-directional LSTM networks

Other LSTM architectures

Exercises

Mining sentiment from Twitter

Connecting to the Twitter API

Building our model

Exploratory data analysis

Using a trained model

Summary

Other Books You May Enjoy

Leave a review - let other readers know what you think