Intelligent Data Engineering and Automated Learning – IDEAL 2022 : 23rd International Conference, IDEAL 2022, Manchester, UK, November 24–26, 2022, Proceedings

This book constitutes the refereed proceedings of the 23rd International Conference on Intelligent Data Engineering and Automated Learning, IDEAL 2022, which took place in Manchester, UK, during November 24-26, 2022. The 52 full papers included in this book were carefully reviewed and selected from 79 submissions. They deal with emerging and challenging topics in intelligent data analytics and associated machine learning paradigms and systems. Special sessions were held on clustering for interpretable machine learning; machine learning towards smarter multimodal systems; and computational intelligence for computer vision and image processing. Preface Organization Contents Main Track Ensemble Stack Architecture for Lungs Segmentation from X-ray Images 1 Introduction 2 Related Works 3 Methodology 4 Experiment 4.1 Evaluation Protocols 4.2 Dataset 4.3 Training Regime 4.4 Results 5 Comparison with State-of-the-Arts 6 Conclusion References Synonym-Based Essay Generation and Augmentation for Robust Automatic Essay Scoring 1 Introduction 2 Related Work 3 Proposed Methodology 3.1 Synonym Replacement and Essay Generation 3.2 Data Augmentation 4 Scoring Models 5 Experiment 5.1 Data Sets 5.2 Essay Pre-processing 5.3 Evaluation Methodology 6 Results and Discussion 6.1 Improving Robustness with Adversarial Data Augmentation and Training 7 Conclusions References Characterizing Cardiovascular Risk Through Unsupervised and Interpretable Techniques 1 Introduction 2 Materials and Methods 2.1 Dataset Description 2.2 Unsupervised and Interpretable Methods 3 Experimental Results 3.1 Characterization of Clusters and CVD Risk Analysis 4 Conclusions References Identification of Sedimentary Strata by Segmentation Neural Networks of Oblique Photogrammetry of UAVs 1 Introduction 2 Theoretical Foundations and Related Works 3 Data and Methods 3.1 Segmentation Architecture 3.2 Dataset 4 Experiment and Discussion 4.1 Experiment 4.2 Discussion 5 Conclusion References Detection of False Information in Spanish Using Machine Learning Techniques 1 Introduction 2 Background and Related Work 3 Data and Resources 4 Methodology 4.1 Linguistic Features 4.2 The Conceptual Architecture of the Fine-Tuned Model 4.3 The Technological Implementation 4.4 Evaluation Metrics 5 Results 6 Conclusions and Future Work References An Approach to Authenticity Speech Validation Through Facial Recognition and Artificial Intelligence Techniques 1 Introduction 2 State of Data 2.1 Deception Detection Techniques 2.2 Face Recognition and Face Features Extraction 3 Experiment 3.1 Framework 3.2 Dataset 3.3 Concept Proof 3.4 Training Details 4 Results and Discussion 4.1 Dataset Analysis 4.2 RNN Model 5 Conclusion References Federating Unlabeled Samples: A Semi-supervised Collaborative Framework for Whole Slide Image Analysis 1 Introduction 2 Methodology 2.1 Problem Formulation 2.2 Federated Model 2.3 Self-trained Student Model 3 Experiments and Results 3.1 Dataset 3.2 Implementation Details 3.3 Validation of the Framework 3.4 Results of the Proposed Framework 4 Conclusion References Automatic Exploration of Domain Knowledge in Healthcare 1 Introduction 2 Background 3 DANKFE – DomAiN Knowledge Based Feature Engineering 4 Case Study: Prediction During COVID-19 Pandemic 4.1 Experimental Results 5 Conclusion References On Studying the Effect of Data Quality on Classification Performances 1 Introduction 2 C1: The Perceived Difficulty of Using a Method According to Experts 3 How Good Is a Repairing (Study of C2 to C5) 3.1 Empirical Setup 3.2 C2: Impact of the Degradation of the Data on Repairing Effectiveness 3.3 C3: Effectiveness of the Repairing Tools 3.4 C4 and C5: Impact of the Type of Error and Impact of the Classification Model 4 Discussion 4.1 Is It Always Better to Repair Data? 4.2 Threats to Validity 5 Conclusion References A Binary Water Flow Optimizer Applied to Feature Selection 1 Introduction 2 Water Flow Optimizer 2.1 Laminar Operator 2.2 Turbulent Operator 2.3 Algorithm 3 Proposal: Binary Water Flow Optimizer (BWFO) 3.1 Binary Laminar Flow Operator 3.2 Binary Turbulent Flow 3.3 Framework BWFO 4 Simulations and Discussions 5 Conclusion References Benchmarking Data Augmentation Techniques for Tabular Data 1 Introduction 2 State of Art 3 Experiments 3.1 Data 3.2 Assessment Metrics 3.3 Experimental Results 4 Conclusion References Deep Learning Based Predictive Analytics for Decentralized Content Caching in Hierarchical Edge Networks 1 Introduction 2 Literature Review 3 Related Works 4 Methodology 4.1 System Architecture 4.2 Dataset Preprocessing 4.3 Model Specification 5 Implementation 5.1 Constructing the Model 5.2 Content Caching and Replacing 6 Result Analysis 7 Conclusion References Explanations of Performance Differences in Segment Lining for Tunnel Boring Machines 1 Introduction 2 Related Work 3 Methods 3.1 Performance Classification 3.2 Model Evaluation 4 Results 4.1 Model Performance Comparison 4.2 Feature Representation Extraction 5 Discussion 6 Conclusion References On Autonomous Drone Navigation Using Deep Learning and an Intelligent Rainbow DQN Agent 1 Introduction 2 Preliminaries 2.1 Value Function 2.2 Multilayer Perceptron Neural Networks 3 Methodology 3.1 Deep Q Networks 3.2 Double Deep Q Networks 3.3 Learning with Multiple Training Cycles 3.4 Rainbow Agent 3.5 Problem Formulation 4 Experimental Results 5 Conclusions and Future Work References An Intelligent Decision Support System for Road Freight Transport 1 Introduction 2 Related Work 3 Materials and Methods 3.1 Problem Formulation 3.2 Proposed IDSS 3.3 Evaluation Methodology 4 Results 4.1 Developed IDSS Prototype 4.2 Evaluation 5 Conclusions References Endowing Intelligent Vehicles with the Ability to Learn User's Habits and Preferences with Machine Learning Methods 1 Introduction 2 Overview of Applied Techniques 2.1 Clustering Approaches for Point of Interest (POI) Extraction 2.2 Artificial Neural Networks 2.3 Regressions 3 Methodology 3.1 Predicting the Next Vehicle Trip State 3.2 Predicting the Comfort Setting 4 Results 4.1 Datasets 4.2 Next Trip State of a Vehicle 4.3 Next Trip's Comfort Setting 5 Conclusion References Duplication Scheduling with Bottom-Up Top-Down Recursive Neural Network 1 Introduction 2 Preliminaries 2.1 Recursive Neural Network 2.2 Bottom-Up Top-Down Recursive Neural Network 3 Experiments 4 Results and Performance Comparison 5 Conclusion References Towards a Low-Cost Companion Robot for Helping Elderly Well-Being 1 Introduction 2 System Description 2.1 Hardware Description 2.2 Software Description 3 Conclusions and Future Works References Zero-Shot Knowledge Graph Completion for Recommendation System 1 Introduction 2 Related Work 3 Our Approach 3.1 Framework 3.2 Problem Formulation 3.3 Zero-Shot KGC 4 Experiments 4.1 Dataset 4.2 Data Pre-processing 4.3 Experimental Setup 4.4 Experiments Result and Comparisons 5 Conclusion and Future Work References The Covid-19 Influence on the Desire to Stay at Home: A Big Data Architecture 1 Introduction 2 State of the Art 3 Materials and Methods 3.1 Sources and Data Collection 3.2 Storage 3.3 ETL Process 3.4 Data Visualization 4 Results and Discussion 4.1 Use Case 1: United States of America 4.2 Use Case 2: India 4.3 Use Case 3: Brazil 5 Conclusions References Distance-Based Delays in Echo State Networks 1 Introduction 2 Methods 3 Results 4 Conclusion 5 Future Work References EduBot: A Proof-of-Concept for a High School Motivational Agent 1 Introduction 2 State of the Art 3 Dataset Presentation 4 An Active Motivational Digital Assistant 4.1 Education Data Mining 4.2 Education Intelligence Module 4.3 Digital Assistant Motivational Module 5 Conclusion References A Simulation Model for Predicting the Spread of COVID-19 Virus 1 Introduction 2 Methods 3 Results 4 Conclusions References ICU Mortality Prediction Using Long Short-Term Memory Networks 1 Introduction 2 Related Works 3 Dataset 3.1 Feature Engineering 3.2 Feature Preprocessing 4 Methodology 4.1 Model Configuration 4.2 Model Implementation 5 Experimental Results 6 Conclusion and Future Works References Intelligent Learning Rate Distribution to Reduce Catastrophic Forgetting in Transformers 1 Introduction 2 Related Work 3 Proposed Approach 3.1 Search Space 3.2 Combining Learning Rate Distributions 4 Experimental Approach 4.1 Datasets 4.2 Types of Data Shift 4.3 Baselines and Implementation Details 5 Results 5.1 Dataset Shift 5.2 Distribution Shift 6 Conclusion References How Image Retrieval and Matching Can Improve Object Localisation on Offshore Platforms 1 Introduction 2 Related Work 3 ARRANGE: ImAge RetRieval mAtchiNG ObjEct 3.1 Principle 3.2 Image Retrieval 3.3 Image Matching 4 Performance Evaluation 4.1 ARRANGE`s Analysis 4.2 ARRANGE Vs. State-of.the-art Object Localisation Algorithms 5 Conclusion References Ethereum Investment Based on LSTM and GRU Forecast 1 Introduction 2 Materials: Data and Pre-processing 2.1 Feature Selection 3 Methods: Neural Networks Application 3.1 Network Architecture and Parametrization 3.2 Metrics 3.3 Training and Testing Data 3.4 Forecasting Results 4 Investment Strategy 5 Conclusions and Future Works References Generating a European Portuguese BERT Based Model Using Content from Arquivo.pt Archive 1 Introduction and Motivation 2 Arquivo.pt 2.1 Arquivo.pt Interfaces 3 Content Retrieve Process 3.1 Step 1: Retrieve Websites 3.2 Step 2: Retrieve Links 3.3 Step 3: Retrieve URL Content 3.4 Step 4: Split Phrases 3.5 Step 5: Model Training 4 Conclusions and Future Work References A Vision Transformer Enhanced with Patch Encoding for Malware Classification 1 Introduction 2 Related Works 3 Proposed Method 3.1 Vision Transformer to Use Location Information of Local Features and Relationship Information 4 Experiments 4.1 Data Collection and Implementation Details 4.2 Quantitative Experimental Results 4.3 Visualization of Activation Area 5 Conclusion and Discussion References Association Rules Mining for Reducing Items from Emotion Regulation Questionnaires 1 Introduction 2 Review on Some Methods Applied for Reducing Items from Questionnaires 3 Association Rules Mining as an Intelligent Approach for Reducing Items 3.1 ERQ Data Set Description 3.2 Association Rules Mining 4 Conclusion References Explainable Artificial Intelligence for Improved Modeling of Processes 1 Introduction 2 Methodology 2.1 Data 2.2 Encoding 2.3 Models 2.4 XAI 3 Experimental Results 3.1 Data Split and Pre-processing 3.2 Evaluation Metric 3.3 Model Design and Training 3.4 Results 4 Conclusion References Efficient Sensor Selection for Individualized Prediction Based on Biosignals 1 Introduction 2 Related Work 3 Methods 3.1 Heuristic Placement Based on Foot Anatomy 3.2 Feature Selection as Global Optimization Task 3.3 Feature Selection as Sparse Approximation 3.4 Dataset 4 Experiments and Results 5 Conclusion References Understanding the Classes Better with Class-Specific and Rule-Specific Feature Selection, and Redundancy Control in a Fuzzy Rule Based Framework 1 Introduction 2 Proposed Method 2.1 Fuzzy Rule-Based Classifiers 2.2 Feature Selection 2.3 Monitoring Redundancy 2.4 Exploiting Substructures Within a Class 3 Experiments and Results 3.1 Experiment 1 3.2 Experiment 2 3.3 Experiment 3 4 Conclusion References Performance/Resources Comparison of Hardware Implementations on Fully Connected Network Inference 1 Introduction 2 Related Work 3 Inference 3.1 Linear Quantized Inference 3.2 UINT8 ONNX Inference 3.3 INT8 Tensorflow Lite Inference 4 Hardware Implementation 4.1 HLS Description 5 Experimental Results 5.1 Accelerated Flow 5.2 Model Architecture and Dataset 5.3 Results and Discussion 6 Conclusions and Future Work References Gradient Regularization with Multivariate Distribution of Previous Knowledge for Continual Learning 1 Introduction 2 Related Works 3 Proposed Method 3.1 Model Architecture 3.2 Training in Initial Task 3.3 Training from the Next Task 4 Experiments 4.1 Datasets and Implementation Details 4.2 Results 4.3 Discussions 5 Concluding Remarks References Face ReID Method via Deep Learning 1 Introduction 1.1 Related Works 1.2 Proposal 2 Methodology 2.1 Database 2.2 Architecture 2.3 Feature Extraction 2.4 Proposed Strategy 2.5 Results and Discussion References Using Design of Experiments to Support the Commissioning of Industrial Assembly Processes 1 Introduction 2 State of the Art 3 Assembled Products 3.1 Formal Description of the DoE Task 3.2 Drum Assembly Use Case 4 Methodology 4.1 Design Optimization 4.2 Incorporation of Angular Positions 4.3 Subset Selection 5 Experiments 5.1 Experimental Design 5.2 Modeling 6 Conclusions and Future Work References Res-GAN: Residual Generative Adversarial Network for Coronary Artery Segmentation 1 Introduction 2 Proposed Method 2.1 Network Architecture 2.2 Objective Loss Function 3 Experiments and Results 3.1 Materials and Experimental Setup 3.2 Evaluation Metrics 3.3 Results 4 Conclusion References Using GANs to Improve the Accuracy of Machine Learning Models for Malware Detection 1 Introduction 2 Related Work 3 Problem Description and Proposed Approach 4 Databases 4.1 Sample Collection 4.2 Feature Mining 4.3 Feature Processing and Selection 5 Experimental Setup 6 Results 7 Conclusion and Future Work References Randomized K-FACs: Speeding Up K-FAC with Randomized Numerical Linear Algebra 1 Introduction 2 Preliminaries 2.1 Fisher Information, Natural Gradient and K-FAC 2.2 Randomized SVD (RSVD) 2.3 Symmetric Randomized EVD (SREVD) 3 The Decaying Eigen-Spectrum of K-Factors 4 Speeding Up EA K-Factors Inversion 4.1 Proposed Optimizer: RSVD K-FAC (RS-KFAC) 4.2 Proposed Optimizer: SREVD K-FAC (SRE-KFAC) 4.3 Direct Idea Transfer to Other Applications 5 Numerical Results: Proposed Algorithms Performance 6 Conclusion References Guide-Guard: Off-Target Predicting in CRISPR Applications 1 Introduction 2 Background and Related Work 2.1 CRISPR 3 A General Model for Gene Editing 4 Methodology 4.1 Mismatch Location 4.2 Nucleotide Replaced 4.3 Guide-Guard 4.4 Data Preparation 4.5 Network Design 5 Results and Discussion 6 Conclusion References Topological Analysis of Credit Data: Preliminary Findings 1 Introduction 2 Literature Review: Credit Scoring 3 Concepts in Topological Data Analysis 3.1 The Vietoris-Rips Filtration 3.2 Barcode 3.3 Landscape 4 Application Pipeline to Credit Data 4.1 Homologies for Neighbourhoods 4.2 Detailed LR+TDA Credit Risk algorithm 5 Results 5.1 Data 5.2 Exploratory Classification Analysis 5.3 Sampling Results 6 Discussion References A Comparative Study of LAD, CNN and DNN for Detecting Intrusions 1 Introduction 2 Related Work 3 Logical Analysis of Data and Its Implementation 3.1 Binarization 3.2 Support Set Generation 3.3 Pattern Generation 3.4 Classifier Design and Validation 4 Performance Evaluation 4.1 Datasets 4.2 Experimental Setup 4.3 Experimental Results for UNSW-NB15 Dataset 4.4 Experimental Results for CSE-CIC-IDS2018 Dataset 5 Conclusion References Effective Prevention of Semantic Drift in Continual Deep Learning 1 Introduction 2 Related Work 3 Proposed Approach 4 Experimental Setup and Results 5 Conclusion References A Sequence to Sequence Long Short-Term Memory Network for Footwear Sales Forecasting 1 Introduction 2 Related Work 3 Materials and Methods 3.1 Footwear Sales Data 3.2 Modeling 3.3 Evaluation 4 Results 5 Conclusions References EfficientNet Architecture Family Analysis on Railway Track Defects 1 Introduction 2 Methodology 2.1 Model Set-Up 3 Dataset Analysis and Pre-processing 4 Results 4.1 Fault Detection 5 Conclusions and Future Work References Challenging Mitosis Detection Algorithms: Global Labels Allow Centroid Localization 1 Introduction 2 Related Work 2.1 Mitosis Detection 2.2 Weakly Supervised Segmentation 3 Methods 4 Experimental Setting 4.1 Datasets 4.2 Metrics 4.3 Implementation Details 5 Results 5.1 Comparison to Literature 5.2 Ablation Experiments 5.3 Qualitative Evaluation 6 Conclusions References Go-Around Prediction in Non-Stabilized Approach Scenarios Through a Regression Machine-Learning Model Trained from Pilots' Expertise 1 Introduction 2 Materials 2.1 Data Acquisition 2.2 Data Description 2.3 Data Conditioning and Pre-processing 3 Methodology 3.1 Go-Around AI Prediction Module 4 Results 4.1 Training Learning Curves 4.2 Model Prediction Performance 5 Discussion 6 Conclusion References Special Session on Intelligent Techniques for Real-World Applications of Renewable Energy and Green Transport Identification of Variables of a Floating Wind Turbine Prototype 1 Introduction 2 Small Replica and Experiment Set up 3 Experimental Data 4 Supervised Regression Models 5 Results 6 Conclusions and Future Works References Dynamic Optimization of Energy Hubs with Evolutionary Algorithms Using Adaptive Time Segments and Varying Resolution 1 Introduction 2 Related Work 3 Method 3.1 Identifying Time Segments for Adaptive Time Resolution 3.2 Combining Schedules 4 Evaluation 4.1 Specification of Evaluation Environment and Criteria 4.2 Results 5 Conclusion and Outlook References Special Session on Computational Intelligence for Imbalanced Classification Solving Multi-class Imbalance Problems Using Improved Tabular GANs 1 Introduction 2 Multi-class Imbalance Learning 2.1 Handling Multi-class Imbalanced Problems 2.2 Data Augmentation Using Tabular GANs 3 Methodology 3.1 Data Pre-processing 3.2 Data Augmentation 3.3 Proposed Data Filter 3.4 Classification Using Decision Trees 4 Experimental Results 4.1 Datasets Description 4.2 Evaluation Metrics 4.3 Classification Results 5 Conclusion References Convolutional Neural Network Approach for Multiple Sclerosis Lesion Segmentation 1 Introduction 2 Multiple Sclerosis Data Base 3 Proposed Automatic MS Lesion Segmentation Method 3.1 Preprocessing 3.2 Patch Extraction 3.3 DL Segmentation Architectures 3.4 Post-processing 4 Results and Evaluation Metrics 4.1 Implementation Details 4.2 Results on ISBI Database 5 Discussion and Conclusion References Author Index