
Artificial Neural Networks in Pattern Recognition
Description
Alles über E-Books | Antworten auf Fragen rund um E-Books, Kopierschutz und Dateiformate finden Sie in unserem Info- & Hilfebereich.
The 29 revised full papers presented together with 2 invited papers were carefully reviewed and selected from 35 submissions. The papers present and discuss the latest research in all areas of neural network- and machine learning-based pattern recognition. They are organized in two sections: learning algorithms and architectures, and applications.
Chapter "Bounded Rational Decision-Making with Adaptive Neural Network Priors" is available open access under a Creative Commons Attribution 4.0 International License via link.springer.com.
More details
Other editions
Additional editions

Content
- Intro
- Preface
- Organization
- Contents
- Invited Papers
- What's Wrong with Computer Vision?
- 1 Introduction
- 2 Top Ten Questions a Theory on Vision Should Address
- 3 Hierarchical Description of Visual Tasks
- 3.1 Pixel-Wise and Abstract Visual Interpretations
- 3.2 The Interwound Story of Vision and Language
- 3.3 When Vision Collapses to Classification
- 4 Conclusions
- References
- Deep Learning in the Wild
- 1 Introduction
- 2 Face Matching
- 3 Print Media Monitoring
- 4 Visual Quality Control
- 5 Music Scanning
- 6 Game Playing
- 7 Automated Machine Learning
- 8 Conclusions
- References
- Learning Algorithms and Architectures
- Effect of Equality Constraints to Unconstrained Large Margin Distribution Machines
- 1 Introduction
- 2 Least Squares Support Vector Machines
- 3 Large Margin Distribution Machines and Their Variants
- 3.1 Large Margin Distribution Machines
- 3.2 Least Squares Large Margin Distribution Machines
- 3.3 Unconstrained Large Margin Distribution Machines
- 4 Performance Evaluation
- 4.1 Conditions for Experiment
- 4.2 Results for Two-Class Problems
- 5 Conclusions
- References
- DLL: A Fast Deep Neural Network Library
- 1 Introduction
- 2 DLL: Deep Learning Library
- 2.1 Performance
- 2.2 Example
- 3 Experimental Evaluation
- 4 MNIST
- 4.1 Fully-Connected Neural Network
- 4.2 Convolutional Neural Network
- 5 CIFAR-10
- 6 ImageNet
- 7 Conclusion and Future Work
- References
- Selecting Features from Foreign Classes
- 1 Introduction
- 2 Methods
- 2.1 Learning from Context Classes
- 2.2 Foreign Class Combinations
- 3 Experiments
- 3.1 Datasets
- 4 Results
- 5 Discussion and Conclusion
- References
- A Refinement Algorithm for Deep Learning via Error-Driven Propagation of Target Outputs
- 1 Introduction
- 2 Error-Driven Target Propagation: Formalization of the Algorithms
- 2.1 The Inversion Net
- 2.2 Refinement of Deep Learning via Target Propagation
- 3 Experiments
- 4 Conclusions
- References
- Combining Deep Learning and Symbolic Processing for Extracting Knowledge from Raw Text
- 1 Introduction
- 2 Model
- 2.1 Semantic Features
- 2.2 Logic Constraints
- 2.3 Segmentation
- 3 Experiments
- 4 Conclusions
- References
- SeNA-CNN: Overcoming Catastrophic Forgetting in Convolutional Neural Networks by Selective Network Augmentation
- 1 Introduction
- 2 Related Work
- 3 Proposed Method
- 4 Experiments
- 4.1 Network Architecture
- 4.2 Training Methodology
- 4.3 Isolated Learning
- 4.4 Adding New Tasks to the Models
- 4.5 Three Tasks Scenario
- 5 Conclusion
- References
- Classification Uncertainty of Deep Neural Networks Based on Gradient Information
- 1 Introduction
- 2 Entropy, Softmax Baseline and Gradient Metrics
- 3 Meta Classification - A Benchmark Between Maximum Softmax Probability and Gradient Metrics
- 4 Recognition of Unlearned Concepts
- 5 Meta Classification with Known Unknowns
- 6 Conclusion and Outlook
- References
- Learning Neural Models for End-to-End Clustering
- 1 Introduction
- 2 Related Work
- 3 A Model for End-to-End Clustering of Arbitrary Data
- 3.1 Network Architecture
- 3.2 Training and Loss
- 3.3 Implicit vs. Explicit Distance Learning
- 4 Experimental Results
- 5 Discussion and Conclusions
- References
- A k-Nearest Neighbor Based Algorithm for Multi-Instance Multi-Label Active Learning
- 1 Introduction
- 2 Method
- 2.1 MIML Framework
- 2.2 MIML-kNN
- 2.3 Active Learning
- 3 Experiments
- 4 Conclusion
- References
- Manifold Learning Regression with Non-stationary Kernels
- 1 Introduction
- 1.1 Nonlinear Multi-output Regression
- 1.2 Learning with Non-stationary Kernels
- 1.3 Manifold Learning Regression
- 1.4 Paper Contribution
- 2 Manifold Learning Regression
- 2.1 Tangent Bundle Manifold Estimation Problem
- 2.2 Grassmann and Stiefel Eigenmaps Algorithm
- 2.3 Manifold Learning Regression Algorithm
- 3 Modified Manifold Learning Regression
- 3.1 Preliminary Estimation of Unknown Functions
- 3.2 Estimation of Jacobian Matrix at Arbitrary Point
- 3.3 Estimation of Unknown Function at Arbitrary Point
- 4 Conclusion
- References
- F-Measure Curves for Visualizing Classifier Performance with Imbalanced Data
- 1 Introduction
- 2 Performance Metrics and Visualization Tools
- 3 The F-Measure Space
- 4 Synthetic Examples
- 5 Conclusions
- References
- Maximum-Likelihood Estimation of Neural Mixture Densities: Model, Algorithm, and Preliminary Experimental Evaluation
- 1 Introduction
- 2 Model and Estimation Algorithm
- 3 Preliminary Experimental Evaluation
- 4 Pro Tempore Conclusions and On-Going Research
- References
- Generating Bounding Box Supervision for Semantic Segmentation with Deep Learning
- 1 Introduction
- 2 Related Works
- 3 The BBSDL Method
- 4 Experiments and Results
- 4.1 The Datasets
- 4.2 Weak Supervision Generation for Pascal-VOC 2012
- 4.3 Experimental Results
- 5 Conclusions and Future Perspectives
- References
- Inductive-Transductive Learning with Graph Neural Networks
- 1 Introduction
- 2 The Graph Neural Network Model
- 3 Transductive Learning with GNNs
- 4 Experimental Evaluation
- 4.1 Datasets
- 4.2 Experimental Setup
- 4.3 Results
- 5 Conclusions
- References
- Bounded Rational Decision-Making with Adaptive Neural Network Priors
- 1 Introduction
- 2 Methods
- 2.1 Bounded Rational Decision Making
- 2.2 MCMC as Sample-Based Bounded Rational Decision-Making
- 2.3 Representing Prior Strategies with Variational Autoencoders
- 3 Modeling Bounded Rationality with Adaptive Neural Network Priors
- 3.1 Decision Making with Multiple Priors
- 3.2 Model Architecture
- 4 Empirical Results
- 5 Discussion
- References
- Feature Selection with Rényi Min-Entropy
- 1 Introduction
- 2 Preliminaries
- 3 Our Proposed Algorithm
- 3.1 An Example in Which Rényi Min-Entropy Gives a Better Feature Selection Than Shannon Entropy
- 3.2 An Example in Which Shannon Entropy May Give a Better Feature Selection Than Rényi Min-Entropy
- 4 Evaluation
- 4.1 Experiments
- 5 Related Works
- 6 Conclusion and Future Work
- References
- Applications
- Extracting Gamma-Ray Information from Images with Convolutional Neural Network Methods on Simulated Cherenkov Telescope Array Data
- 1 Introduction
- 2 Monte Carlo Simulation and Preselection
- 3 Convolutional Neural Networks for Simulated Cherenkov Telescope Array Data
- 4 Summary and Conclusion
- References
- Automatic Hand Sign Recognition: Identify Unusuality Through Latent Cognizance
- 1 Introduction
- 2 Background
- 3 Prediction Confidence and Non-sign Identification
- 4 Experiments
- 5 Discussion and Conclusions
- References
- Cascade of Ordinal Classification and Local Regression for Audio-Based Affect Estimation
- Abstract
- 1 Introduction
- 2 State of Art
- 2.1 Emotion Classification and Prediction
- 2.2 Ordinal Classification and Hierarchical Prediction
- 3 Cascade of Ordinal Classifiers and Local Regressors
- 3.1 Ordinal Classification
- 3.2 Local Regression
- 4 Databases
- 4.1 AVEC'2014
- 4.2 AV+EC'2015/RECOLA
- 4.3 Data Augmentation
- 5 Experimental Results
- 5.1 Performance Metrics
- 5.2 Preliminary Results
- 5.3 Results on AVEC'2014
- 5.4 Results on AV+EC'2015/ RECOLA
- 5.5 Temporal Smoothing
- 6 Conclusions
- References
- Combining Classical and Deep Learning Methods for Twitter Sentiment Analysis
- 1 Introduction
- 2 The Previous Work
- 3 Theory and Algorithms
- 3.1 Data Processing
- 3.2 Auto-Labeling
- 3.3 Feature Extraction
- 3.4 Classifiers
- 4 The Proposed Models
- 4.1 Model 1: BOW with tf-idf
- 4.2 Model 2: Aggregation of Word2Vec
- 4.3 Model 3: LSTM
- 4.4 Model 4: CNN
- 4.5 Model 5: The Proposed Model
- 5 Experiments
- 5.1 The Datasets
- 5.2 The Experimental Protocol
- 5.3 The Experimental Results
- 6 Conclusions and Future Work
- References
- PHoG Features and Kullback-Leibler Divergence Based Ranking Method for Handwriting Recognition
- 1 Introduction
- 2 Related Work
- 3 The Proposed System
- 3.1 Preprocessing
- 3.2 Decision Trees Construction
- 3.3 The Ranking Stage
- 4 Experiments
- 4.1 Decision Trees Sensitivity
- 4.2 The Ranking Process
- 4.3 Complexity Analysis
- 4.4 Comparison with Similar Systems
- 5 Conclusions
- References
- Pattern Recognition Pipeline for Neuroimaging Data
- 1 Introduction
- 2 MR Data Properties, Preprocessing and Feature Extraction
- 2.1 Structural MRI Preprocessing and Feature Extraction
- 2.2 Functional MRI Preprocessing and Feature Extraction
- 3 The Proposed Pipeline
- 4 Illustrative Example: Pattern Recognition for Epilepsy Detection
- 4.1 fMRI-based Pattern Recognition
- 4.2 Structural MRI-based Pattern Recognition
- 5 Conclusions
- References
- Anomaly Pattern Recognition with Privileged Information for Sensor Fault Detection
- 1 Introduction
- 2 Anomaly Detection
- 2.1 One Class SVM
- 2.2 One Class SVM+
- 3 Sensor Fault Detection in Road Weather Information Systems
- 3.1 Sensor Data
- 3.2 METRo Model
- 3.3 Learning Sample
- 3.4 Anomaly Detection Accuracy Metric
- 3.5 One Class SVM Parameters
- 3.6 Results
- 4 Conclusions
- References
- Capturing Suprasegmental Features of a Voice with RNNs for Improved Speaker Clustering
- 1 Introduction
- 2 Speaker Clustering with RNNs
- 2.1 Related Work
- 2.2 Overview of Our Approach
- 3 Experimental Evaluation
- 3.1 Experimental Setup
- 3.2 Results
- 4 Discussion
- 5 Conclusions and Future Work
- References
- Trace and Detect Adversarial Attacks on CNNs Using Feature Response Maps
- 1 Introduction
- 2 Related Work
- 3 Background
- 3.1 Adversarial Attacks
- 3.2 Feature Response Estimation
- 4 Human-Interpretable Detection of Adversarial Attacks
- 4.1 Tracing Adversarial Attacks in Feature Responses
- 4.2 Detecting Adversarial Attacks Using Spatial Entropy
- 5 Experimental Results
- 6 Discussion and Conclusion
- References
- Video and Audio Data Extraction for Retrieval, Ranking and Recapitulation (VADER3)
- 1 Introduction
- 2 Approach
- 2.1 Action Recognition
- 2.2 Voiceover Detection
- 2.3 Speech Recognition
- 2.4 Automated Scene Captioning
- 2.5 Text Detection and Recognition (OCR) and Object Recognition
- 2.6 Language Model Based Video Similarity
- 2.7 Video Ranking and Retrieval
- 2.8 Clustering
- 2.9 Semantic Labeling
- 3 Experimental Evaluation
- 3.1 Dataset
- 3.2 Clustering and Labeling
- 3.3 Individual Modules
- 4 Conclusion
- References
- ATM Protection Using Embedded Deep Learning Solutions
- 1 Introduction
- 2 Related Works
- 2.1 Video Surveillance and Action Recognition
- 2.2 ATM
- 3 ATMSense
- 3.1 Depth Cameras
- 3.2 Image Pre-processing
- 3.3 Deep Neural Networks
- 4 Experiments
- 5 Conclusions
- 6 Future Works
- References
- Object Detection in Floor Plan Images
- 1 Introduction
- 1.1 Previous Work
- 2 The Architecture
- 2.1 False Negative Calculation
- 3 The Floor Plan Datasets
- 4 Experiments
- 4.1 Discussion
- 5 Conclusions
- References
- Historical Handwritten Document Segmentation by Using a Weighted Loss
- 1 Introduction
- 2 Related Works
- 3 Network Architecture
- 4 Weighting the Loss
- 4.1 Weighting Background Pixels
- 5 Experiments
- 6 Conclusions
- References
- Author Index
System requirements
File format: PDF
Copy protection: Watermark-DRM (Digital Rights Management)
System requirements:
- Computer (Windows; MacOS X; Linux): Use the free software Adobe Reader, Adobe Digital Editions, or any other PDF viewer of your choice (see eBook Help).
- Tablet/Smartphone (Android; iOS): Install the free app Adobe Digital Editions or another reading app for eBooks, e.g., PocketBook (see eBook Help).
- E-reader: Bookeen, Kobo, Pocketbook, Sony, Tolino and many more (only limited: Kindle).
The file format PDF always displays a book page identically on any hardware. This makes PDF suitable for complex layouts such as those used in textbooks and reference books (images, tables, columns, footnotes). Unfortunately, on the small screens of e-readers or smartphones, PDFs are rather annoying, requiring too much scrolling.
This eBook uses Watermark-DRM, a „soft” copy protection. This means that there are no technical restrictions to prevent illegal distribution. However, there is a personalised watermark embedded in the eBook that can be used to identify the purchaser of the eBook in the event of misuse and to provide evidence for legal purposes.
For more information, see our eBook Help page.