Paperback(1st Edition)
-
PICK UP IN STORECheck Availability at Nearby Stores
Available within 2 business hours
Related collections and offers
Overview
Deep Learning with R introduces the world of deep learning using the powerful Keras library and its R language interface. The book builds your understanding of deep learning through intuitive explanations and practical examples.
Continue your journey into the world of deep learning with Deep Learning with R in Motion, a practical, hands-on video course available exclusively at Manning.com (www.manning.com/livevideo/deep-learning-with-r-in-motion).
Purchase of the print book includes a free eBook in PDF, Kindle, and ePub formats from Manning Publications.
About the Technology
Machine learning has made remarkable progress in recent years. Deep-learning systems now enable previously impossible smart applications, revolutionizing image recognition and natural-language processing, and identifying complex patterns in data. The Keras deep-learning library provides data scientists and developers working in R a state-of-the-art toolset for tackling deep-learning tasks.
About the Book
Deep Learning with R introduces the world of deep learning using the powerful Keras library and its R language interface. Initially written for Python as Deep Learning with Python by Keras creator and Google AI researcher François Chollet and adapted for R by RStudio founder J. J. Allaire, this book builds your understanding of deep learning through intuitive explanations and practical examples. You'll practice your new skills with R-based applications in computer vision, natural-language processing, and generative models.
What's Inside
- Deep learning from first principles
- Setting up your own deep-learning environment
- Image classification and generation
- Deep learning for text and sequences
About the Reader
You'll need intermediate R programming skills. No previous experience with machine learning or deep learning is assumed.
About the Authors
François Chollet is a deep-learning researcher at Google and the author of the Keras library.
J.J. Allaire is the founder of RStudio and the author of the R interfaces to TensorFlow and Keras.
Table of Contents
- What is deep learning?
- Before we begin: the mathematical building blocks of neural networks
- Getting started with neural networks
- Fundamentals of machine learning
- Deep learning for computer vision
- Deep learning for text and sequences
- Advanced deep-learning best practices
- Generative deep learning
- Conclusions
Product Details
| ISBN-13: | 9781617295546 |
|---|---|
| Publisher: | Manning |
| Publication date: | 02/09/2018 |
| Edition description: | 1st Edition |
| Pages: | 360 |
| Sales rank: | 1,131,660 |
| Product dimensions: | 7.30(w) x 9.10(h) x 0.90(d) |
About the Author
J.J. Allaire is the Founder of RStudio and the creator of the RStudio IDE. J.J. is the author of the R interfaces to TensorFlow and Keras.
Table of Contents
Preface xiii
Acknowledgments xv
About this book xvi
About the authors xx
About the cover xxi
Part 1 Fundamentals of Deep Learning 1
1 What is deep learning? 3
1.1 Artificial intelligence, machine learning, and deep learning 4
Artificial intelligence 4
Machine learning 4
Learning representations from data 6
The "deep" in deep learning 8
Understanding how deep learning works, in three figures 9
What deep learning has achieved so far 11
Don't believe the short-tern hype 12
The promise of AI 12
1.2 Before deep learning: a brief history of machine learning 13
Probabilistic modeling 14
Early neural networks 14
Kernel methods 15
Decision trees, random forests, and gradient boosting machines 16
Back to neural networks 17
What makes deep learning different 17
The modern machine-learning landscape 18
1.3 Why deep learning? Why now? 19
Hardware 19
Data 20
Algorithms 21
A new wave of investment 21
The democratization of deep learning 22
Will it last? 22
2 Before we begin: the mathematical building blocks of neural networks 24
2.1 A first look at a neural network 25
2.2 Data representations for neural networks 29
Scalars (0D tensors) 29
Vectors (1D tensors) 29
Matrices (2D tensors) 30
3D tensors and higher-dimensional tensors 30
Key attributes 30
Manipulating tensors in R 31
The notion of data batches 32
Real-world examples of data tensors 32
Vector data 32
Timeseries data or sequence data 33
Image data 33
Video data 34
2.3 The gears of neural networks: tensor operations 34
Element-wise operations 35
Operations involving tensors of different dimensions 36
Tensor dot 36
Tensor reshaping 38
Geometric interpretation of tensor operations 39
A geometric interpretation of deep learning 40
2.4 The engine of neural networks: gradient-based optimization 41
What's a derivative? 42
Derivative of a tensor operation: the gradient 43
Stochastic gradient descent 44
Chaining derivatives: the Backpropagation algorithm 46
2.5 Looking back at our first example 47
2.6 Summary 49
3 Getting started with neural networks 50
3.1 Anatomy of a neural network 51
Layers: the building blocks of deep learning 52
Models: networks of layers 52
Loss functions and optimizers: keys to configuring the learning process 53
3.2 Introduction to Keras 54
Kerns, TensorFlow, Theano, and CNTK 54
Installing Keras 56
Developing with Kerns: a quick overview 56
3.3 Setting up a deep-learning workstation 57
Getting Keras running: two options 58
Running deep-learning jobs in the cloud: pros and cons 58
What is the best GPU for deep learning? 59
3.4 Classifying movie reviews: a binary classification example 59
The IMDB dataset 59
Preparing the data 61
Building your network 62
Validating your approach 65
Using a trained network to generate predictions on new data 68
Further experiments 69
Wrapping up 69
3.5 Classifying newswires: a multiclass classification example 70
The Reuters dataset 70
Preparing the data 71
Building your network 72
Validating your approach 73
Generating predictions on new data 74
A different way to handle the labels and the loss 75
The importance of having sufficiently large intermediate layers 75
Further experiments 76
Wrapping up 76
3.6 Predicting house prices: a regression example 76
The Boston Housing Price dataset 77
Preparing the data 77
Building your network 78
Validating your approach using K-fold validation 79
Wrapping up 83
3.7 Summary 83
4 Fundamentals of machine learning 84
4.1 Four branches of machine learning 85
Supervised learning 85
Unsupervised learning 85
Self-supervised learning 86
Reinforcement learning 86
4.2 Evaluating machine-learning models 87
Training, validation, and lest sets 88
Things to keep in mind 91
4.3 Data preprocessing, feature engineering, and feature learning 91
Data preprocessing for neural networks 91
Feature engineering 93
4.4 Overfitting and underfitting 94
Reducing the network's size 95
Adding weight regularization 98
Adding dropout 100
4.5 The universal workflow of machine learning 102
Defining the problem and assembling a dataset 102
Choosing a measure of success 103
Deciding on an evaluation protocol 104
Preparing your data 104
Developing a model that does better than a baseline 104
Scaling up: developing a model that overfits 105
Regularizing your model and tuning your hyperparameters 106
4.6 Summary 107
Part 2 Deep Learning in Practice 109
5 Deep learning for computer vision 111
5.1 Introduction to convnets 111
The convolution operation 114
The max-pooling operation 119
5.2 Training a convnet from scratch on a small dataset 121
The relevance of deep learning for small-data problems 121
Downloading the data 122
Building your network 124
Data preprocessing 126
Using data augmentation 128
5.3 Using a pretrained convnet 132
Feature extraction 133
Fine-tuning 142
Wrapping up 146
5.4 Visualizing what convnets learn 146
Visualizing intermediate activations 146
Visualizing convnet filters 153
Visualizing heatmaps of class activation 159
5.5 Summary 163
6 Deep learning for text and sequences 164
6.1 Working with text data 165
One-hot encoding of words and characters 166
Using word embeddings 169
Putting it all together: from raw text to word embeddings 174
Wrapping up 180
6.2 Understanding recurrent neural networks 180
A recurrent layer in Keras 182
Understanding the LSTM and GRU layers 186
A concrete LSTM example in Keras 188
Wrapping up 190
6.3 Advanced use of recurrent neural networks 190
A temperature-forecasting problem 191
Preparing the data 193
A common-sense, non-machine-learning baseline 197
A basic machine-learning approach 198
A first recurrent baseline 199
Using recurrent dropout to fight overfilling 201
Stacking recurrent layers 202
Using bidirectional RNNs 204
Going even further 207
Wrapping up 208
6.4 Sequence processing with convnets 209
Understanding ID convolution for sequence data 209
ID pooling for sequence data 210
Implementing a ID convnet 210
Combining CNNs and RNNs to process long sequences 212
Wrapping up 216
6.5 Summary 216
7 Advanced deep-learning best practices 218
7.1 Going beyond the sequential model: the Keras functional API 219
Introduction to the functional API 221
Multi-input models 222
Multi-output models 224
Directed acyclic graphs of layers 227
Layer weight sharing 231
Models as layers 232
Wrapping up 233
7.2 Inspecting and monitoring deep-learning models using Keras callbacks and TensorBoard 233
Using callbacks to act on a model during training 233
Introduction to TensorBoard: the TensorFlow visualization framework 236
Wrapping up 241
7.3 Getting the most out of your models 241
Advanced architecture patterns 241
Hyperparameter optimization 245
Model ensembling 246
Wrapping up 248
7.4 Summary 249
8 Generative deep learning 250
8.1 Text generation with LSTM 252
A brief history of generative recurrent networks 252
How do you generate sequence data? 253
The importance of the sampling strategy 253
Implementing character-level LSTM text generation 255
Wrapping up 260
8.2 DeepDream 260
Implementing DeepDream in Keras 261
Wrapping up 267
8.3 Neural style transfer 267
The content loss 268
The style loss 268
Neural style transfer in Keras 269
Wrapping up 274
8.4 Generating images with variational autoencoders 276
Sampling from latent spaces of images 276
Concept vectors for image editing 277
Variational autoencoders 278
Wrapping up 284
8.5 Introduction to generative adversarial networks 284
A schematic GAN implementation 286
A bag of tricks 286
The generator 287
The discriminator 288
The adversarial network 289
How to train your DCGAN 290
Wrapping up 292
8.6 Summary 292
9 Conclusions 293
9.1 Key concepts in review 294
Various approaches to Al 294
What makes deep learning special within the field of machine learning 294
How to think about deep learning 295
Key enabling technologies 296
The universal machine-learning workflow 297
Key network architectures 298
The space of possibilities 302
9.2 The limitations of deep learning 303
The risk of anthropomorphizing machine-learning models 304
Local generalization vs. extreme generalization 306
Wrapping up 307
9.3 The future of deep learning 307
Models as programs 308
Beyond backpropagation and differentiable layers 310
Automated machine learning 310
Lifelong learning and modular subroutine reuse 311
The long-term vision 313
9.4 Staying up to date in a fast-moving field 313
Practice on real-world problems using Kaggle 314
Read about the latest developments on arXiv 314
Explore the Keras ecosystem 315
9.5 Final words 315
Appendix A Installing Kerns and its dependencies on Ubuntu 316
Appendix B Running RStudio Server on an EC2 GPU instance 320
Index 327