Vieira / Ribeiro Introduction to Deep Learning Business Applications for Developers

From Conversational Bots in Customer Service to Medical Image Processing
1. ed
ISBN: 978-1-4842-3453-2
Verlag: Apress
Format: PDF
Kopierschutz: 1 - PDF Watermark

E-Book, Englisch, 348 Seiten

Discover the potential applications, challenges, and opportunities of deep learning from a business perspective with technical examples. These applications include image recognition, segmentation and annotation, video processing and annotation, voice recognition, intelligent personal assistants, automated translation, and autonomous vehicles. 
An Introduction to Deep Learning Business Applications for Developers covers some common DL algorithms such as content-based recommendation algorithms and natural language processing. You'll explore examples, such as video prediction with fully convolutional neural networks (FCNN) and residual neural networks (ResNets). You will also see applications of DL for controlling robotics, exploring the DeepQ learning algorithm with Monte Carlo Tree search (used to beat humans in the game of Go), and modeling for financial risk assessment. There will also be mention of the powerful set of algorithms called Generative Adversarial Neural networks (GANs) that can be applied for image colorization, image completion, and style transfer.
After reading this book you will have an overview of the exciting field of deep neural networks and an understanding of most of the major applications of deep learning. The book contains some coding examples, tricks, and insights on how to train deep learning models using the Keras framework.
What You Will Learn
Find out about deep learning and why it is so powerful
Work with the major algorithms available to train deep learning models
See the major breakthroughs in terms of applications of deep learning  
Run simple examples with a selection of deep learning libraries 
Discover the areas of impact of deep learning in business

Who This Book Is For Data scientists, entrepreneurs, and business developers.


Dr Armando Vieira is a Data Scientist and Artificial Intelligence consultant with an entrepreneurial mindset. Passionate about how to make Machine Learning projects work for organizations and how to build great AI based products.As algorithms are becoming a commodity, the challenge is not building them but using them to solve real problems.
Have coordinated several projects on Credit Risk Evaluation, Recommendation Systems, Clustering Analysis and Predictive Analytics. 

Bernardete Ribeiro is Professor at University of Coimbra, Portugal. She has a Ph.D. and Habilitation in Informatics Engineering. She is Director of the Center of Informatics and Systems of the University of Coimbra (CISUC).She is President of the Portuguese Association of Pattern Recognition (APRP). She is Founder and Director of the Laboratory of Artificial Neural Networks (LARN) for more than 20 years. She is IEEE SMC Senior member, member of International Association of Pattern Recognition (IAPR), International Neural Network Society (INNS), and ACM. Her research interests are in the areas of Machine Learning, Pattern Recognition, and their applications to abroad range of fields. She is author or co-author of over three hundred publications including books, journalsand international and national conferences. She has delivered numerous invited talks, seminars, and short courses.

Autoren/Hrsg.

Vieira, Armando

Ribeiro, Bernardete

Inhaltsverzeichnis

1;Table of Contents;5
2;About the Authors;12
3;About the Technical Reviewer;13
4;Acknowledgments;15
5;Introduction;16
6;Part I: Background and Fundamentals;19
6.1;Chapter 1: Introduction;20
6.1.1;1.1 Scope and Motivation;21
6.1.2;1.2 Challenges in the Deep Learning Field;23
6.1.3;1.3 Target Audience;23
6.1.4;1.4 Plan and Organization;24
6.2;Chapter 2: Deep Learning: An Overview;25
6.2.1;2.1 From a Long Winter to a Blossoming Spring;27
6.2.2;2.2 Why Is DL Different?;30
6.2.2.1;2.2.1 The Age of the Machines;33
6.2.2.2;2.2.2 Some Criticism of DL;34
6.2.3;2.3 Resources;35
6.2.3.1;2.3.1 Books;35
6.2.3.2;2.3.2 Newsletters;36
6.2.3.3;2.3.3 Blogs;36
6.2.3.4;2.3.4 Online Videos and Courses;37
6.2.3.5;2.3.5 Podcasts;38
6.2.3.6;2.3.6 Other Web Resources;39
6.2.3.7;2.3.7 Some Nice Places to Start Playing;40
6.2.3.8;2.3.8 Conferences;41
6.2.3.9;2.3.9 Other Resources;42
6.2.3.10;2.3.10 DL Frameworks;42
6.2.3.11;2.3.11 DL As a Service;45
6.2.4;2.4 Recent Developments;48
6.2.4.1;2.4.1 2016;48
6.2.4.2;2.4.2 2017;49
6.2.4.3;2.4.3 Evolution Algorithms;50
6.2.4.4;2.4.4 Creativity;51
6.3;Chapter 3: Deep Neural Network Models;52
6.3.1;3.1 A Brief History of Neural Networks;53
6.3.1.1;3.1.1 The Multilayer Perceptron;55
6.3.2;3.2 What Are Deep Neural Networks?;57
6.3.3;3.3 Boltzmann Machines;60
6.3.3.1;3.3.1 Restricted Boltzmann Machines;63
6.3.3.1.1;Contrastive Divergence;64
6.3.3.2;3.3.2 Deep Belief Nets;65
6.3.3.3;3.3.3 Deep Boltzmann Machines;68
6.3.4;3.4 Convolutional Neural Networks;69
6.3.5;3.5 Deep Auto-encoders;70
6.3.6;3.6 Recurrent Neural Networks;71
6.3.6.1;3.6.1 RNNs for Reinforcement Learning;74
6.3.6.2;3.6.2 LSTMs;76
6.3.7;3.7 Generative Models;79
6.3.7.1;3.7.1 Variational Auto-encoders;80
6.3.7.2;3.7.2 Generative Adversarial Networks;84
7;Part II: Deep Learning: Core Applications;89
7.1;Chapter 4: Image Processing;90
7.1.1;4.1 CNN Models for Image Processing;91
7.1.2;4.2 ImageNet and Beyond;94
7.1.3;4.3 Image Segmentation;99
7.1.4;4.4 Image Captioning;102
7.1.5;4.5 Visual Q&A (VQA);103
7.1.6;4.6 Video Analysis;107
7.1.7;4.7 GANs and Generative Models;111
7.1.8;4.8 Other Applications;115
7.1.8.1;4.8.1 Satellite Images;116
7.1.9;4.9 News and Companies;118
7.1.10;4.10 Third-Party Tools and APIs;121
7.2;Chapter 5: Natural Language Processing and Speech;123
7.2.1;5.1 Parsing;125
7.2.2;5.2 Distributed Representations;126
7.2.3;5.3 Knowledge Representation and Graphs;128
7.2.4;5.4 Natural Language Translation;135
7.2.5;5.5 Other Applications;139
7.2.6;5.6 Multimodal Learning and Q&A;141
7.2.7;5.7 Speech Recognition;142
7.2.8;5.8 News and Resources;145
7.2.9;5.9 Summary and a Speculative Outlook;148
7.3;Chapter 6: Reinforcement Learning and Robotics;149
7.3.1;6.1 What Is Reinforcement Learning?;150
7.3.2;6.2 Traditional RL;152
7.3.3;6.3 DNN for Reinforcement Learning;154
7.3.3.1;6.3.1 Deterministic Policy Gradient;155
7.3.3.2;6.3.2 Deep Deterministic Policy Gradient;155
7.3.3.3;6.3.3 Deep Q-learning;156
7.3.3.4;6.3.4 Actor-Critic Algorithm;159
7.3.4;6.4 Robotics and Control;162
7.3.5;6.5 Self-Driving Cars;165
7.3.6;6.6 Conversational Bots (Chatbots);167
7.3.7;6.7 News Chatbots;171
7.3.8;6.8 Applications;173
7.3.9;6.9 Outlook and Future Perspectives;174
7.3.10;6.10 News About Self-Driving Cars;176
8;Part III: Deep Learning: Business Applications;181
8.1;Chapter 7: Recommendation Algorithms and E-commerce;182
8.1.1;7.1 Online User Behavior;183
8.1.2;7.2 Retargeting;184
8.1.3;7.3 Recommendation Algorithms;186
8.1.3.1;7.3.1 Collaborative Filters;187
8.1.3.2;7.3.2 Deep Learning Approaches to RSs;189
8.1.3.3;7.3.3 Item2Vec;191
8.1.4;7.4 Applications of Recommendation Algorithms;192
8.1.5;7.5 Future Directions;193
8.2;Chapter 8: Games and Art;196
8.2.1;8.1 The Early Steps in Chess;196
8.2.2;8.2 From Chess to Go;197
8.2.3;8.3 Other Games and News;199
8.2.3.1;8.3.1 Doom;199
8.2.3.2;8.3.2 Dota;199
8.2.3.3;8.3.3 Other Applications;200
8.2.4;8.4 Artificial Characters;202
8.2.5;8.5 Applications in Art;203
8.2.6;8.6 Music;206
8.2.7;8.7 Multimodal Learning;208
8.2.8;8.8 Other Applications;209
8.3;Chapter 9: Other Applications;217
8.3.1;9.1 Anomaly Detection and Fraud;218
8.3.1.1;9.1.1 Fraud Prevention;221
8.3.1.2;9.1.2 Fraud in Online Reviews;223
8.3.2;9.2 Security and Prevention;224
8.3.3;9.3 Forecasting;226
8.3.3.1;9.3.1 Trading and Hedge Funds;228
8.3.4;9.4 Medicine and Biomedical;231
8.3.4.1;9.4.1 Image Processing Medical Images;232
8.3.4.2;9.4.2 Omics;235
8.3.4.3;9.4.3 Drug Discovery;238
8.3.5;9.5 Other Applications;240
8.3.5.1;9.5.1 User Experience;240
8.3.5.2;9.5.2 Big Data;241
8.3.6;9.6 The Future;242
9;Part IV: Opportunities and Perspectives;244
9.1;Chapter 10: Business Impact of DL Technology;245
9.1.1;10.1 Deep Learning Opportunity;247
9.1.2;10.2 Computer Vision;248
9.1.3;10.3 AI Assistants;249
9.1.4;10.4 Legal;251
9.1.5;10.5 Radiology and Medical Imagery;252
9.1.6;10.6 Self-Driving Cars;254
9.1.7;10.7 Data Centers;255
9.1.8;10.8 Building a Competitive Advantage with DL;255
9.1.9;10.9 Talent;257
9.1.10;10.10 It’s Not Only About Accuracy;259
9.1.11;10.11 Risks;260
9.1.12;10.12 When Personal Assistants Become Better Than Us;261
9.2;Chapter 11: New Research and Future Directions;263
9.2.1;11.1 Research;264
9.2.1.1;11.1.1 Attention;265
9.2.1.2;11.1.2 Multimodal Learning;266
9.2.1.3;11.1.3 One-Shot Learning;267
9.2.1.4;11.1.4 Reinforcement Learning and Reasoning;269
9.2.1.5;11.1.5 Generative Neural Networks;271
9.2.1.6;11.1.6 Generative Adversarial Neural Networks;272
9.2.1.7;11.1.7 Knowledge Transfer and Learning How to Learn;274
9.2.2;11.2 When Not to Use Deep Learning;276
9.2.3;11.3 News;277
9.2.4;11.4 Ethics and Implications of AI in Society;279
9.2.5;11.5 Privacy and Public Policy in AI;282
9.2.6;11.6 Startups and VC Investment;284
9.2.7;11.7 The Future;287
9.2.7.1;11.7.1 Learning with Less Data;289
9.2.7.2;11.7.2 Transfer Learning;290
9.2.7.3;11.7.3 Multitask Learning;290
9.2.7.4;11.7.4 Adversarial Learning;291
9.2.7.5;11.7.5 Few-Shot Learning;291
9.2.7.6;11.7.6 Metalearning;292
9.2.7.7;11.7.7 Neural Reasoning;292
10;Appendix A:Training DNN with Keras;294
10.1;A.1 The Keras Framework;294
10.1.1;A.1.1 Installing Keras in Linux;295
10.1.2;A.1.2 Model;295
10.1.3;A.1.3 The Core Layers;296
10.1.4;A.1.4 The Loss Function;298
10.1.5;A.1.5 Training and Testing;298
10.1.6;A.1.6 Callbacks;299
10.1.7;A.1.7 Compile and Fit;299
10.2;A.2 The Deep and Wide Model;300
10.3;A.3 An FCN for Image Segmentation;310
10.3.1;A.3.1 Sequence to Sequence;314
10.4;A.4 The Backpropagation on a Multilayer Perceptron;317
11;References;325
12;Index;338