Title Page Copyright and Credits Advanced Deep Learning with R About Packt Why subscribe? Contributors About the author About the reviewer 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 Conventions used Get in touch Reviews Section 1: Revisiting Deep Learning Basics Revisiting Deep Learning Architecture and Techniques Deep learning with R Deep learning trend Versions of key R packages used Process of developing a deep network model Preparing the data for a deep network model Developing a deep learning model architecture Compiling the model Fitting the model Assessing the model performance Deep learning techniques with R and RStudio Multi-class classification Regression problems Image classification Convolutional neural networks Autoencoders Transfer learning Generative adversarial networks Deep network for text classification  Recurrent neural networks  Long short-term memory network Convolutional recurrent networks Tips, tricks, and best practices Summary Section 2: Deep Learning for Prediction and Classification Deep Neural Networks for Multi-Class Classification Cardiotocogram dataset Dataset (medical) Preparing the data for model building Normalizing numeric variables Partitioning the data One-hot encoding Creating and fitting a deep neural network model Developing model architecture Compiling the model Fitting the model Model evaluation and predictions Loss and accuracy calculation Confusion matrix Performance optimization tips and best practices Experimenting with an additional hidden layer Experimenting with a higher number of units in the hidden layer Experimenting using a deeper network with more units in the hidden layer Experimenting by addressing the class imbalance problem Saving and reloading a model Summary Deep Neural Networks for Regression Understanding the Boston Housing dataset Preparing the data Visualizing the neural network Data partitioning Normalization Creating and fitting a deep neural network model for regression Calculating the total number of parameters Compiling the model Fitting the model Model evaluation and prediction Evaluation Prediction Improvements Deeper network architecture Results Performance optimization tips and best practices Log transformation on the output variable Model performance Summary Section 3: Deep Learning for Computer Vision Image Classification and Recognition Handling image data Data preparation Resizing and reshaping Training, validation, and test data One-hot encoding Creating and fitting the model Developing the model architecture Compiling the model Fitting the model Model evaluation and prediction Loss, accuracy, and confusion matrices for training data Prediction probabilities for training data Loss, accuracy, and confusion matrices for test data Prediction probabilities for test data Performance optimization tips and best practices Deeper networks Results Summary Image Classification Using Convolutional Neural Networks Data preparation Fashion-MNIST data Train and test data Reshaping and resizing One-hot encoding Layers in the convolutional neural networks Model architecture and related calculations Compiling the model Fitting the model Accuracy and loss Model evaluation and prediction Training data Test data 20 fashion items from the internet Performance optimization tips and best practices Image modification Changes to the architecture Summary Applying Autoencoder Neural Networks Using Keras Types of autoencoders Dimension reduction autoencoders MNIST fashion data Encoder model Decoder model Autoencoder model Compiling and fitting the model Reconstructed images Denoising autoencoders MNIST data Data preparation Adding noise Encoder model Decoder model Autoencoder model Fitting the model Image reconstruction Image correction Images that need correction Clean images Encoder model Decoder model Compiling and fitting the model Reconstructing images from training data Reconstructing images from new data Summary Image Classification for Small Data Using Transfer Learning Using a pretrained model to identify an image Reading an image Preprocessing the input Top five categories Working with the CIFAR10 dataset Sample images Preprocessing and prediction Image classification with CNN Data preparation CNN model Model performance Performance assessment with training data Performance assessment with test data Classifying images using the pretrained RESNET50 model Model architecture Freezing pretrained network weights Fitting the model Model evaluation and prediction Loss, accuracy, and confusion matrix with the training data Loss, accuracy, and confusion matrix with the test data Performance optimization tips and best practices Experimenting with the adam optimizer Hyperparameter tuning Experimenting with VGG16 as a pretrained network Summary Creating New Images Using Generative Adversarial Networks Generative adversarial network overview Processing MNIST image data Digit five from the training data Data processing Developing the generator network Network architecture Summary of the generator network Developing the discriminator network Architecture Summary of the discriminator network Training the network Initial setup for saving fake images and loss values Training process Reviewing results Discriminator and GAN losses Fake images Performance optimization tips and best practices Changes in the generator and discriminator network Impact of these changes on the results Generating a handwritten image of digit eight Summary Section 4: Deep Learning for Natural Language Processing Deep Networks for Text Classification Text datasets The UCI machine learning repository Text data within Keras Preparing the data for model building Tokenization Converting text into sequences of integers Padding and truncation Developing a tweet sentiment classification model Developing deep neural networks Obtaining IMDb movie review data Building a classification model Compiling the model Fitting the model Model evaluation and prediction Evaluation using training data Evaluation using test data Performance optimization tips and best practices Experimenting with the maximum sequence length and the optimizer Summary Text Classification Using Recurrent Neural Networks Preparing data for model building Padding sequences Developing a recurrent neural network model Calculation of parameters Compiling the model Fitting the model Accuracy and loss Model evaluation and prediction Training the data Testing the data Performance optimization tips and best practices Number of units in the simple RNN layer Using different activation functions in the simple RNN layer Adding more recurrent layers  The maximum length for padding sequences Summary Text classification Using Long Short-Term Memory Network Why do we use LSTM networks? Preparing text data for model building Creating a long short-term memory network model LSTM network architecture Compiling the LSTM network model Fitting the LSTM model Loss and accuracy plot Evaluating model performance  Model evaluation with train data Model evaluation with test data Performance optimization tips and best practices Experimenting with the Adam optimizer Experimenting with the LSTM network having an additional layer Experimenting with a bidirectional LSTM layer Summary Text Classification Using Convolutional Recurrent Neural Networks Working with the reuter_50_50 dataset Reading the training data Reading the test data Preparing the data for model building Tokenization and converting text into a sequence of integers Changing labels into integers Padding and truncation of sequences Data partitioning One-hot encoding the labels Developing the model architecture Compiling and fitting the model Compiling the model Fitting the model Evaluating the model and predicting classes Model evaluation with training data Model evaluation with test data Performance optimization tips and best practices Experimenting with reduced batch size Experimenting with batch size, kernel size, and filters in CNNs Summary Section 5: The Road Ahead Tips, Tricks, and the Road Ahead TensorBoard for training performance visualization Visualizing deep network models with LIME Visualizing model training with tfruns Early stopping of network training Summary Other Books You May Enjoy Leave a review - let other readers know what you think