TY - BOOK AU - Muhamad Amin,Anang Hudaya AU - Khan,Asad I. AU - Nasution,Benny B. TI - Internet-scale pattern recognition: new techniques for voluminous data sets and data clouds SN - 9781466510968 (hardback) U1 - 006.312 23 PY - 2013/// CY - Boca Raton PB - CRC Press KW - Data mining KW - Pattern recognition systems KW - Web usage mining KW - Mathematical statistics KW - Data processing KW - COMPUTERS / Database Management / Data Mining KW - COMPUTERS / Machine Theory KW - COMPUTERS / Internet / General N1 - Includes bibliographical references (p. 167-175) and index; Preface -- Acknowledgments -- About the Authors -- I. Recognition: A New Perspective -- 1. Introduction -- 1.1. As We See, We Learn -- 1.2. Recognition at a Large Scale -- 1.3. Computational Intelligence Approach for Pattern Recognition -- 1.4. Scalability in Pattern Recognition -- 1.4.1. Common Barriers -- 1.4.2. Possible Solutions -- 1.4.3. Distributed Computing Solution for Scalability of PR Schemes -- 2. Distributed Approach for Pattern Recognition -- 2.1. Scalability of Neural Network Approaches -- 2.1.1. Pattern Storage Capacity -- 2.1.2. Inter-Neuron Communication Frequency -- 2.2. Key Components of DPR -- 2.2.1. Learning Mechanism -- 2.2.2. Processing Approach -- 2.2.3. Training Procedure -- 2.3. System Approaches -- 2.4. Pattern Distribution Techniques -- 2.4.1. Subpattern Distribution -- 2.4.2. Pattern Set Distribution -- 2.5. Current DPR Schemes -- 2.5.1. Graph Neuron -- 2.5.2. Hierarchical Graph Neuron -- 2.5.3. Distributed Hierarchical Graph Neuron -- 2.6. Resource Considerations for DPR Implementations -- 2.6.1. Resource-Aware Approach -- 2.6.2. Message-Passing Model in DPR -- II. Evolution of Internet-Scale Recognition -- 3. One-Shot Learning Considerations -- 3.1. One-Shot Learning Graph Neuron (GN) Scheme -- 3.1.1. Pattern Representation -- 3.1.2. Recognition Procedure -- 3.2. One-Shot Learning Model -- 3.2.1. Bias Array Design for Pattern Memorization -- 3.2.2. Collaborative-Comparison Learning Technique -- 3.3. GN Complexity Estimation -- 3.4. Graph Neuron Limitations -- 3.5. Significance of One-Shot Learning -- 4. Hierarchical Model for Pattern Recognition -- 4.1. Evolution of One-Shot Learning: The Hierarchical Approach -- 4.1.1. Solution to Crosstalk Problem -- 4.1.2. Computational Design for a Hierarchical One-Shot Learning DPR Scheme -- 4.1.3. HGN Recognition Procedure -- 4.2. Complexity and Scalability of Hierarchical DPR Scheme -- 4.2.1. Complexity Estimation -- 4.2.2. Scalability in HGN Approach -- 4.3. Reducing Hierarchical Complexity: A Distributed Approach -- 4.3.1. Distributed Neurons of HGN Network -- 4.3.2. Distributed HGN Approach -- 4.4. Design Evaluation for Distributed DPR Approach -- 4.4.1. Non-Uniform Distribution -- 4.4.2. Uniform Distribution -- 5. Recognition via Divide-and-Distribute Approach -- 5.1. Divide-and-Distribute Approach for One-Shot Learning S-PR Scheme -- 5.1.1. Associative Memory (AM) Concept in Pattern Recognition -- 5.1.2. DHGN Computational Design -- 5.1.3. Dual-Phase Recognition Procedure -- 5.2. Dimensionality Reduction in Pattern Pre-Processing -- 5.2.1. Structural Reduction -- 5.2.2. Content Reduction -- 5.3. Remarks on DHGN DPR Scheme -- III. Systems and Tools -- 6. Internet-Scale Applications Development -- 6.1. Distributed Computing Models for IS-PR -- 6.1.1. Commodity Grid (CoG) -- 6.1.2. Cloud Computing -- 6.1.3. Peer-to-Peer (P2P) Computing -- 6.2. Parallel Programming Techniques -- 6.2.1. Message-Passing Scheme -- 6.2.2. GPU Programming -- 6.3. From Coding to Applications -- IV. Implementations and Applications -- 7. Multi-Feature Classifications for Complex Data -- 7.1. Data Features for Pattern Recognition -- 7.2. Distributed Multi-Feature Recognition -- 7.2.1. Conceptual Design and Implementation -- 7.2.2. Complexity Estimation -- 7.3. Handwritten Object Classification with Multiple Features -- 7.3.1. Handwritten Object -- 7.3.2. Classification Procedures -- 7.4. Distributed Multi-Feature Recognition Perspective -- 8. Pattern Recognition within Coarse-Grained Networks -- 8.1. Network Granularity Considerations -- 8.1.1. DHGN Configurations for Adaptive Granularity -- 8.1.2. DHGN Commodity Grid Framework -- 8.2. Face Recognition Using the Multi-Feature DPR Approach -- 8.2.1. Color and Spatio-Structural Features Consideration -- 8.3. Distributed Data Management within Cloud Computing -- 8.3.1. Cloud Data Access Scheme -- 8.3.2. DHGN Approach for Cloud Data Access -- 8.4. Adaptive Recognition: A Different Perspective -- 9. Event Detection within Pine-Grained Networks -- 9.1. Distributed Event Detection Scheme for Wireless Sensor Networks -- 9.1.1. WSN Event Detection -- 9.1.2. DHGN-WSN Event Detection Configuration -- 9.1.3. Dimensionality Reduction in Sensory Data -- 9.1.4. Event Classification -- 9.1.5. Performance Metrics: Memory Utilization -- 9.1.6. Spatio-Temporal Analysis of Event Data -- 9.2. Integrated Grid-Sensor Scheme for Structural Analysis -- 9.2.1. Integrated Grid-Sensor Network Framework for Structural Engineering -- 9.2.2. Structural Analysis, Design, and Monitoring Applications -- 9.3. Distributed Event Detection: A Lightweight Approach -- V. The Way Forward -- 10. Recognition: The Future and Beyond -- 10.1. Medium of Change -- 10.2. Future of Internet-Scale PR -- 10.3. Making a Case -- 10.3.1. Changing the Fundamentals -- 10.3.2. Recognition as Commodity -- Bibliography -- Index N2 - "This cutting-edge reference outlines the underlying theory and principles of efficient and effective distributed pattern recognition involving one-shot learning and in-network processing for different types of applications, including multimedia retrieval systems and event detection over different network environments. Investigating one-shot learning and in-network processing as complementary mechanisms for efficient and accurate distributed pattern analyses, it presents the technical aspects related to the development of scalable pattern recognition using a number of contemporary application development tools. It also considers scalability of pattern recognition schemes when dealing with such data"-- ER -