Deep Learning Algorithms

This book covers different topics from deep learning algorithms, methods and approaches for deep learning, deep learning applications in biology, deep learning applications in medicine, and deep learning applications in pattern recognition systems. Section 1 focuses on methods and approaches for deep learning, describing advancements in deep learning theory and applications - perspective in 2020 and beyond; deep ensemble reinforcement learning with multiple deep deterministic policy gradient algorithm; dynamic decision-making for stabilized deep learning software platforms; deep learning for hyperspectral data classification through exponential momentum deep convolution neural networks; and ensemble network architecture for deep reinforcement learning. Section 2 focuses on deep learning applications in biology, describing fish detection using deep learning; deep learning identification of tomato leaf disease; deep learning for plant identification in natural environment; and applying deep learning models to mouse behavior recognition. Section 3 focuses on deep learning applications in medicine, describing application of deep learning in brain hemorrhage classification using transfer learning; a review of the application of deep learning in brachytherapy; exploring deep learning and transfer learning for colonic polyp classification; and deep learning algorithm for brain-computer interface. Section 4 focuses on deep learning applications in pattern recognition systems, describing application of deep learning in airport visibility forecast; hierarchical representations feature deep learning for face recognition; review of research on text sentiment analysis based on deep learning; classifying hand written digits with deep learning; and bitcoin price prediction based on deep learning methods. Cover 1 Title Page 5 Copyright 6 DECLARATION 7 ABOUT THE EDITOR 9 TABLE OF CONTENTS 11 List of Contributors 17 List of Abbreviations 23 Preface 25 Section 1: Methods and Approaches for Deep Learning 27 Chapter 1 Advancements in Deep Learning Theory and Applications: Perspective in 2020 and Beyond 29 Abstract 29 Introduction 30 Deep Network Topologies 34 Application of Deep Learning 37 Modern Deep Learning Platforms 40 Training Algorithms 43 Routine Challenges of Deep Learning 45 Available Open-Source Datasets 47 References 50 Chapter 2 Deep Ensemble Reinforcement Learning With Multiple Deep Deterministic Policy Gradient Algorithm 55 Abstract 55 Introduction 56 Background 58 Methods 60 Results and Discussion 65 Conclusions 76 References 77 Chapter 3 Dynamic Decision-Making For Stabilized Deep Learning Software Platforms 81 Abstract 81 Introduction 82 Stabilized Control for Reliable Deep Learning Platforms 83 The Use of Lyapunov Optimization for Deep Learning Platforms 89 Emerging Applications 94 Conclusions 95 Acknowledgements 96 References 97 Chapter 4 Deep Learning For Hyperspectral Data Classification Through Exponential Momentum Deep Convolution Neural Networks 99 Abstract 99 Introduction 100 Feature Learning 101 Structure Design of Hyperspectral Data Classification Framework 102 Exponential Momentum Gradient Descent Algorithm 103 Experiment and Analysis 106 Conclusion 112 Acknowledgments 113 References 114 Chapter 5 Ensemble Network Architecture for Deep Reinforcement Learning 119 Abstract 119 Introduction 120 Related Work 121 Ensemble Methods for Deep Reinforcement Learning 123 Experiments 126 Conclusion 128 References 130 Section 2: Deep Learning Techniques Applied in Biology 133 Chapter 6 Fish Detection Using Deep Learning 135 Abstract 135 Introduction 136 Literature Review 137 Materials and Methods 139 Data Augmentation 144 Results and Discussion 152 Conclusion 155 Acknowledgments 156 References 157 Chapter 7 Can Deep Learning Identify Tomato Leaf Disease? 161 Abstract 161 Introduction 162 Related Work 163 Materials and Methods 164 Experiments and Results 169 Conclusion 175 Acknowledgments 176 References 177 Chapter 8 Deep Learning For Plant Identification In Natural Environment 183 Abstract 183 Introduction 184 Proposed Bjfu100 Dataset and Deep Learning Model 185 Experiments and Results 188 Resnet26 on Flavia Dataset 191 Conclusion 192 Acknowledgments 193 References 194 Chapter 9 Applying Deep Learning Models to Mouse Behavior Recognition 197 Abstract 197 Introduction 198 The Mouse Behavior Dataset 200 Experiments and Results 201 Conclusions 212 Acknowledgements 212 References 213 Section 3: Deep learning Applications in Medicine 215 Chapter 10 Application of Deep Learning in Neuroradiology: Brain Hemorrhage Classification Using Transfer Learning 217 Abstract 217 Introduction 218 Related Work 220 Convolutional Neural Network 221 Transfer Learning 222 Materials and Methods 223 Results and Discussion 230 Limitations 236 Conclusion 237 References 238 Chapter 11 A Review of the Application of Deep Learning in Brachytherapy 243 Abstract 243 Introduction 244 Organ Delineation and Segmentation 245 Segmentation and Reconstruction of the Applicator (Interstitial Needles) 246 Dose Calculation 248 Application of Treatment Planning System 248 Others 249 Conclusions 250 References 251 Chapter 12 Exploring Deep Learning and Transfer Learning for Colonic Polyp Classification 255 Abstract 255 Introduction 256 Materials and Methods 258 Results and Discussion 268 Conclusion 276 Acknowledgments 277 References 278 Chapter 13 Deep Learning Algorithm For Brain-Computer Interface 285 Abstract 285 Introduction 286 Critical Review of the Related Literature 299 Comparison of Classification Algorithms 302 Discussion 303 Methodology 306 Conclusion 307 References 308 Section 4: Deep Learning in Pattern Recognition Tasks 311 Chapter 14 The Application of Deep Learning In Airport Visibility Forecast 313 Abstract 313 Introduction 314 Deep Learning 314 The Establishment of Prediction Model 315 Predictive Effect Test 317 Conclusions 321 References 323 Chapter 15 Hierarchical Representations Feature Deep Learning For Face Recognition 325 Abstract 325 Introduction 326 Images Preprocessing 328 Feature Extraction 330 Designing the Classifiers of Supervised Learning 333 Designing the Classifier Combining Unsupervised and Supervised Learning 341 Experiments 348 Conclusion 358 Acknowledgements 358 References 360 Chapter 16 Review of Research on Text Sentiment Analysis Based on Deep Learning 367 Abstract 367 Introduction 368 Brief Review on the Research Progress of Text Sentiment Analysis 369 Introduction to Text Sentiment Analysis Based on Deep Learning 370 Summary and Prospect 374 References 376 Chapter 17 Classifying Hand Written Digits With Deep Learning 379 Abstract 379 Introduction 380 Digit Classification with Deep Networks 380 Experiment 386 Conclusions 387 References 390 Chapter 18 Bitcoin Price Prediction Based on Deep Learning Methods 393 Abstract 393 Introduction 394 Dataset Exploration 394 Pre-Processing 395 Models 395 Results 397 Conclusion and Discussion 401 References 402 Index 403 Back Cover 412