Trends in Deep Learning Methodologies: Algorithms, Applications, and Systems (Hybrid Computational Intelligence for Pattern Analysis and Understanding)

__Trends in Deep Learning Methodologies: Algorithms, Applications, and Systems__ covers deep learning approaches such as neural networks, deep belief networks, recurrent neural networks, convolutional neural networks, deep auto-encoder, and deep generative networks, which have emerged as powerful computational models. Chapters elaborate on these models which have shown significant success in dealing with massive data for a large number of applications, given their capacity to extract complex hidden features and learn efficient representation in unsupervised settings. Chapters investigate deep learning-based algorithms in a variety of application, including biomedical and health informatics, computer vision, image processing, and more. In recent years, many powerful algorithms have been developed for matching patterns in data and making predictions about future events. The major advantage of deep learning is to process big data analytics for better analysis and self-adaptive algorithms to handle more data. Deep learning methods can deal with multiple levels of representation in which the system learns to abstract higher level representations of raw data. Earlier, it was a common requirement to have a domain expert to develop a specific model for each specific application, however, recent advancements in representation learning algorithms allow researchers across various subject domains to automatically learn the patterns and representation of the given data for the development of specific models. TRENDS IN DEEP LEARNING METHODOLOGIES Copyright Contributors Preface 1 . An introduction to deep learning applications in biometric recognition 1. Introduction 1.1 Biometric recognition 1.1.1 Overview 1.1.2 Identification versus verification 1.1.3 Expectations from a biometric security system 2. Methods 2.1 Prominent deep learning algorithms 2.1.1 Convolutional neural networks 2.1.2 Recurrent neural networks 2.1.3 Autoencoder network 2.1.4 Siamese neural network 2.2 Deep learning methodology for precise recognition 2.2.1 Hard biometrics 2.2.1.1 Face recognition 2.2.1.2 Fingerprint recognition 2.2.1.3 Palmprint recognition 2.2.1.4 Iris recognition 2.2.1.5 Vein recognition 2.3 Voice recognition 2.4 Gait recognition 2.4.1 Signature recognition 2.4.2 Soft biometrics 2.5 Databases 2.6 Deep learning methodology for spoof protection and template protection 2.6.1 Spoof protection 2.6.2 Template protection 2.6.2.1 Cancelable biometrics 2.6.2.2 Biometric cryptosystems 2.7 Challenges in biometric recognition and security 3. Comparative analysis among different modalities 4. Further advancement 5. Conclusion References 2 . Deep learning in big data and data mining 1. Introduction 2. Overview of big data analysis 2.1 Variety of data 2.2 Velocity of data 2.3 Volume of data 2.4 Veracity of data 2.5 Variability of data 2.6 Visualization of data 2.7 Value of data 2.8 Distributed computing 2.8.1 MapReduce Framework 2.8.2 Hive 2.8.3 Apache Spark 2.8.3.1 Visualizations using tableau 2.9 Data warehouse versus data lake 3. Introduction 3.1 What is data mining? 3.2 Why use deep learning in data mining? 4. Applications of deep learning in data mining 4.1 Multimedia data mining 4.2 Aspect extraction 4.3 Loss function 4.4 Customer relationship management 5. Conclusion References Further readings 3 . An overview of deep learning in big data, image, and signal processing in the modern digital age 1. Introduction 1.1 Deep learning concepts 1.2 Machine learning 1.3 Big data 1.4 Scientific review 2. Discussion 3. Conclusions 4. Future trends References 4 . Predicting retweet class using deep learning 1. Introduction 2. Related work and proposed work 2.1 Handcrafted feature: basic machine learning classifiers 2.2 Generative models 2.3 Lexical semantics 2.4 Proposed framework 3. Data collection and preparation 3.1 Tweet corpus creation 3.2 Tweet data cleaning 3.3 Tweet data preparation 3.4 Word expletive, a proposed feature 4. Research set-up and experimentation 4.1 The long short-term memory neural network 5. Results 6. Discussion 7. Conclusion References 5 . Role of the Internet of Things and deep learning for the growth of healthcare technology 1. Introduction to the Internet of Things 2. Role of IoT in the healthcare sector 3. IoT architecture 4. Role of deep learning in IoT 5. Design of IoT for a hospital 6. Security features considered while designing and implementing IoT for healthcare 7. Advantages and limitations of IoT for healthcare technology 8. Discussions, conclusions, and future scope of IoT References 6 . Deep learning methodology proposal for the classification of erythrocytes and leukocytes 1. Introduction 2. Hematology background 3. Deep learning concepts 4. Convolutional neural network 5. Scientific review 6. Methodology proposal 7. Results and discussion 8. Conclusions 9. Future research directions References 7 . Dementia detection using the deep convolution neural network method 1. Introduction 2. Related work 3. Basics of a convolution neural network 3.1 Convolution layer 3.2 Pooling layer 3.3 Fully connected layer 4. Materials and methods 4.1 Dataset description 4.2 Deep Learning methodology for dementia detection 5. Experimental results 6. Conclusion References 8 . Deep similarity learning for disease prediction 1. Introduction 2. State of the art 3. Materials and methods 3.1 Data gathering 3.2 Data preprocessing 3.3 Splitting the data 3.4 Model training 3.4.1 Representation learning 3.4.2 Similarity learning 3.5 Evaluation and prediction 4. Results and discussion 5. Conclusions and future work References 9 . Changing the outlook of security and privacy with approaches to deep learning 1. Introduction 2. Birth and history of deep learning 3. Frameworks of deep learning 4. Statistics behind deep learning algorithms and neural networks 5. Deep learning algorithms for securing networks 6. Performance measures for intrusion detection systems 7. Security aspects changing with deep learning 8. Conclusion and future work References 10 . E-CART: an improved data stream mining approach 1. Introduction 2. Related study 3. E-CART: proposed approach 4. Experiment 5. Conclusion References 11 . Deep learning-based detection and classification of adenocarcinoma cell nuclei 1. Introduction 2. Basics of a convolution neural network 2.1 Convolution layer 2.2 Pooling layer 2.3 Fully connected layer 3. Literature review 4. Proposed system architecture and methodology 4.1 Cell detection using faster R-CNN 4.2 Cell classification with ResNet-101 5. Experimentation 5.1 Dataset 5.2 Discussion on results 6. Conclusion References 12 . Segmentation and classification of hand symbol images using classifiers 1. Introduction 2. Literature review 3. Hand symbol classification mechanism 3.1 Image preprocessing stage 3.2 Segmentation stage 3.2.1 Thresholding methods 3.2.2 Histogram-based image segmentation 3.2.3 Feature extraction and selection for making a feature vector 3.2.4 Types of features 3.3 Classification stage 3.3.1 Classification phases 4. Proposed work 5. Results and discussion 6. Conclusion References Index A B C D E F G H I K L M N O P Q R S T U V W Y Z Trends in Deep Learning Methodologies: Algorithms, Applications, and Systems covers deep learning approaches such as neural networks, deep belief networks, recurrent neural networks, convolutional neural networks, deep auto-encoder, and deep generative networks, which have emerged as powerful computational models. Chapters elaborate on these models which have shown significant success in dealing with massive data for a large number of applications, given their capacity to extract complex hidden features and learn efficient representation in unsupervised settings. Chapters investigate deep learning-based algorithms in a variety of application, including biomedical and health informatics, computer vision, image processing, and more. In recent years, many powerful algorithms have been developed for matching patterns in data and making predictions about future events. The major advantage of deep learning is to process big data analytics for better analysis and self-adaptive algorithms to handle more data. Deep learning methods can deal with multiple levels of representation in which the system learns to abstract higher level representations of raw data. Earlier, it was a common requirement to have a domain expert to develop a specific model for each specific application, however, recent advancements in representation learning algorithms allow researchers across various subject domains to automatically learn the patterns and representation of the given data for the development of specific models. Provides insights into the theory, algorithms, implementation and the application of deep learning techniques Covers a wide range of applications of deep learning across smart healthcare and smart engineering Investigates the development of new models and how they can be exploited to find appropriate solutions