Engineering Applications of Neural Networks: 24th International Conference, EAAAI/EANN 2023, León, Spain, June 14–17, 2023, Proceedings (Communications in Computer and Information Science, 1826)

Preface
Organization
Abstracts of Invited Talks
	Evolutionary Neural Architecture Search: Computational Efficiency, Privacy Preservation and Robustness Enhancement
	Interpretable-by-Design Prototype-Based Deep Learning
	Intelligent Mobile Sensing for Understanding Human Behaviour
	Secure, Efficient and High-Performance Computing: A Computer Architecture Perspective
	How AI/Machine Learning has the Power of Revolutionizing (for Good?) Cybersecurity?
	Contents
Artificial Intelligence - Computational Methods - Ethology
A Machine Learning Approach for Seismic Vulnerability Ranking
	1 Introduction and Literature Review
	2 Dataset
	3 Data Preprocessing - Pairwise Ranking Transformation
	4 Description of the Algorithm
		4.1 General Overview
		4.2 Employed Machine Learning Modelling Algorithms
	5 Results
	6 Conclusions and Future Work
	References
Computational Ethology: Short Review of Current Sensors and Artificial Intelligence Based Methods
	1 Introduction
	2 Sensors
	3 Methods
		3.1 Tracking
		3.2 Behavioral Classification
	4 Conclusion and Future Perspective
	References
Toward an Epidermal Patch for Voice Prosthesis and Diseases Prediction by a Fuzzy-Neuro Paradigm
	1 Introduction
	2 Energy Harvesting
	3 Flexible PCB
	4 Ng and Processing the Myoelectric Signal
	5 Electro Myo GRAM and Phonemes
	6 He EFuNN Paradigm
	7 Daftaset, Training and Test
	8 Conclusion and Future Developments
	References
Towards Improved User Experience for Artificial Intelligence Systems
	1 Introduction
	2 Related Work
	3 OMA-ML User Interaction Concept
	4 UX Evaluation of OMA-ML
		4.1 Methodology
		4.2 Main Results
	5 Recommendations for UX in AI Systems
	6 Conclusions and Future Work
	References
Classification - Filtering - Genetic Algorithms
A Critical Analysis of Classifier Selection in Learned Bloom Filters: The Essentials
	1 Introduction
	2 Bloom Filters and Learned Bloom Filters
	3 Experimental Methodology
		3.1 A Representative Set of Binary Classifiers
		3.2 Measures of Classification Complexity and a Related Data Generation Procedure
	4 Experiments
		4.1 Data
		4.2 Model Selection
	5 Results and Discussion
	6 Guidelines
	7 Conclusions and Future Developments
	References
Balancing High-Dimensional Datasets with Complex Layers
	1 Introduction
	2 High-Dimensional Learning Sets
	3 Linear Separability Resulting from Linear Independence
	4 Perceptron Criterion Function
	5 Vertices in Parameter Space
	6 Optimal Subsets of Features
	7 Balanced Complex Layers
	8 Classifiers Based on Complex Layers
	9 Concluding Remarks
	References
BotDroid: Permission-Based Android Botnet Detection Using Neural Networks
	1 Introduction
	2 Background
		2.1 Overview of Android Botnet
		2.2 Types of Botnets
		2.3 Botnet lifecycle
		2.4 Botnet Attacks
	3 Related works
		3.1 Static Techniques
		3.2 Dynamic Techniques
		3.3 Hybrid Techniques
	4 Dataset
		4.1 Feature Selection
		4.2 Feature Extraction
	5 Proposed Method
	6 Experimental Evaluation
		6.1 Evaluation Metrics
		6.2 Results
		6.3 Comparisons with other Classifiers
		6.4 Comparisons with the Most Recent Botnet Detection Studies
	7 Conclusions
	References
Classification of Time Signals Using Machine Learning Techniques
	1 Introduction
	2 Background
		2.1 Hilbert Transform
	3 Methodology
		3.1 Data Reading
		3.2 Applying Hilbert Transform
		3.3 Support Vector Machine (SVM) Implemention
		3.4 Implementing Multi-Layer Perceptron (MLP)
	4 Final Results of SVM and MLP Model
	References
Conductivity Classification Using Machine Learning Algorithms in the “Bramianon” Dam
	1 Introduction
		1.1 Defining the Problem
		1.2 Aim of This Machine Learning Modeling Effort
		1.3 Literature Review
	2 Area of Research – Data
	3 Machine Learning Modeling
		3.1 Water Conductivity Modeling
		3.2 Development of ML Models
		3.3 Models’ Assessment Indices
	4 Performance Results
	5 Discussion-Conclusion
	References
Generation of Bases for Classification in the Bio-inspired Layered Networks
	1 Introduction
	2 Bio-inspired Neural Networks
		2.1 Background of Asymmetric Neural Networks Based on the Bio-inspired Network
		2.2 Model of Asymmetric Networks
		2.3 Orthogonality and Independence of Bases in the Asymmetric Network Unit
	3 Classification Evaluation in the Asymmetric Network
		3.1 Conditions of independence of \"026B30D  Asym. \"026B30D  and \"026B30D  Sym. \"026B30D
		3.2 Patterns Design for Independence in Asymmetric and Symmetric Networks
		3.3 Generation of Combined Bases for Orthogonality
		3.4 Generation of the Higher Dimensional Bases in the Layered Networks
		3.5 Replacement of the Orthogonal Bases in the 2nd Layered Network
	4 Conclusion
	References
Load-Shedding Management in a Smart Grid Architecture Through Smart Metering
	1 Introduction
		1.1 Smart Grid Architecture and Smart Meter
	2 Previous Work
	3 Methodology
		3.1 Experimental Setup
		3.2 Computational Intelligence
		3.3 Data Collection, Presentation and Intervals
		3.4 Data Normalization and Processing
	4 Results
		4.1 Load-Forecasting Algorithm
		4.2 Load-Balancing Algorithm
		4.3 Performance of the Genetic Algorithm
	5 Discussion
	6 Conclusion and Way Forward
	References
MiniAnDE: A Reduced AnDE Ensemble to Deal with Microarray Data
	1 Introduction
	2 Averaged n-Dependence Estimators (AnDE)
	3 MiniAnDE
	4 Experimental Evaluation
		4.1 Data Sets
		4.2 Reproducibility
		4.3 Algorithms
		4.4 Methodology
		4.5 Results
	5 Conclusions
	References
Multi-view Semi-supervised Learning Using Privileged Information
	1 Introduction
	2 Task of Multi-View Semi-Supervised Learning
	3 Multiple Self Training with Privileged Information
	4 Experimental Study
		4.1 Task of Predicting Student Advice for Secondary Education
		4.2 Experimental Setup
		4.3 Results
	5 Conclusion
	References
Complex Dynamic Networks’ Optimization/Graph Neural Networks
A Multi-relationship Language Acquisition Model for Predicting Child Vocabulary Growth
	1 Introduction
	2 Existing Work
	3 Child Vocabulary as a Multi-relationship Graph
	4 Model Selection
	5 Methodology
		5.1 Assumptions and Data Preparation
		5.2 Training and Validation
		5.3 Ensemble Models
		5.4 Evaluation
	6 Conclusions and Future Work
	References
Knowledge Graph of Urban Firefighting with Rule-Based Entity Extraction
	1 Introduction
	2 Related Work
		2.1 Urban Firefighting Ontology
		2.2 Urban Firefighting Knowledge Graph
	3 A Knowledge Graph for Urban Firefighting
		3.1 The Framework for Building the Urban Firefighting Knowledge Graph
		3.2 Construction of Urban Firefighting Ontology
		3.3 Building the Urban Firefighting Knowledge Graph
	4 Verification of the Proposed Method
		4.1 Data Acquisition and Preprocessing
		4.2 The Knowledge Graph of Urban Firefighting
		4.3 Result and Discussion
	5 Conclusion
	References
Optimal Traffic Flow Distributions on Dynamic Networks
	1 Introduction
	2 The Model of a Network
	3 The Main Results
		3.1 Wardrop Optimal Flows
		3.2 Wardrop Optimal Flows on Dynamic Networks
	4 Multi-agent Intelligent Transport System
	5 Conclusion
	A  Appendix
	References
Streamlined Training of GCN for Node Classification with Automatic Loss Function and Optimizer Selection
	1 Introduction
	2 Proposed Architecture
		2.1 Automatic Search and Selection of Best Functions
		2.2 Real-time Training Visualization
	3 Dataset
	4 Findings and Analysis
	5 Conclusion
	References
User Equilibrium and System Optimality Conditions for Flow Distributions on Congested Networks
	1 Introduction
	2 Preliminaries
	3 User Equilibrium and System Optimum
	4 Wardrop Optimal Networks
	5 Conclusion and Future Work
	References
Convolutional Neural Networks/Spiking Neural Networks
Digital Transformation for Offshore Assets: A Deep Learning Framework for Weld Classification in Remote Visual Inspections
	1 Introduction
	2 Related Work
	3 Methods
		3.1 Data Collection
		3.2 Data Pre-processing
		3.3 Transfer Learning
	4 Experiments and Results
		4.1 Experiment Setup
		4.2 Results
	5 Conclusion and Future Work
	References
Frameworks for SNNs: A Review of Data Science-Oriented Software and an Expansion of SpykeTorch
	1 Introduction
	2 Related Works
	3 Software Frameworks
		3.1 Nengo
		3.2 SNN Toolbox
		3.3 Lava
		3.4 PyTorch-Based Frameworks
	4 SpykeTorch Spiking Neurons
		4.1 Spiking Neurons Implementation Details
		4.2 Heterogeneous Neuron Classes
	5 Conclusions
	References
Hierarchical Prediction in Incomplete Submetering Systems Using a CNN
	1 Introduction
	2 Related Work
	3 The Proposed Method
	4 Experiments and Results
		4.1 Submetering System and Dataset
		4.2 Experiments
		4.3 Results and Discussion
	5 Conclusions
	References
.25em plus .1em minus .1emPruning for Power: Optimizing Energy Efficiency in IoT with Neural Network Pruning*-12pt
	1 Introduction
	2 Theoretical Background
		2.1 TinyML
		2.2 Neural Network Pruning
	3 Experiments
		3.1 Experimental Setup
		3.2 Implementation
	4 Experimental Results
	5 Unstructured Pruning Speedup
		5.1 Performance Bottleneck and Improvements
		5.2 Results
	6 Conclusion
		6.1 Recommendations
		6.2 Delimitation and Further Research
	A Measurement Setup
	B Experimental Results
	References
Deep Learning Modeling
Advanced Skin Cancer Detection Using Deep Learning
	1 Introduction
	2 Literature Review
	3 Methodology
		3.1 Preprocessing and Segmentation
		3.2 CNN Classifiers
		3.3 SVM Classifier with Feature Extraction
	4 Experimental Design
		4.1 Dataset
		4.2 Experiments
	5 Implementation Details
	6 Results and Discussion
		6.1 Parameters Tuning
		6.2 Results
	7 Conclusion and Future Work
	References
Cross-Learning-Based Sales Forecasting Using Deep Learning via Partial Pooling from Multi-level Data*-12pt
	1 Introduction
	2 Deep Learning Model
		2.1 DeepAR
		2.2 Tweedie Loss
	3 Empirical Study
		3.1 Dataset
		3.2 Partial Pooling of Multi-level Data
		3.3 Model Selection
		3.4 Results
	4 Conclusions
	References
Discrimination of Attention Deficit Hyperactivity Disorder Using Capsule Networks and LSTM Networks on fMRI Data
	1 Introduction
	2 Literature Survey
	3 Dataset
	4 Proposed System
		4.1 Preprocessing and Feature Extraction
		4.2 Capsule Networks
		4.3 LSTM Networks
	5 Results and Discussion
	6 Conclusion and Future Scope
	References
Gaussian-Based Approach for Out-of-Distribution Detection in Deep Learning
	1 Introduction
	2 Related Works
	3 Our Proposed Approaches
		3.1 Gaussian Max Likelihood
		3.2 Gaussian Mixture Max Likelihood
	4 Experiments
		4.1 Datasets
		4.2 OOD Methods
		4.3 Metrics
		4.4 Experimental Details
	5 Results and Discussion
		5.1 Image Classification
		5.2 Text Classification
	6 Conclusion
	References
LoockMe: An Ever Evolving Artificial Intelligence Platform for Location Scouting in Greece
	1 Introduction
	2 Material and Methods
		2.1 Data Collection
		2.2 Deep Learning-Based Image Analysis
		2.3 Natural Language Processing Service
		2.4 LoockMe Platform
	3 Results
	4 Discussion
	5 Conclusion
	References
Object Detection for Functional Assessment Applications
	1 Introduction
	2 GEFAD-Objects Dataset
	3 Object Detection for Functional Assessment Evaluation
		3.1 Object Detection by Pattern
		3.2 Object Detection by Colour
	4 Experimental Results
	5 Conclusions and Future Works
	References
Performance Analysis of Digit Recognizer Using Various Machine Learning Algorithms
	1 Introduction
	2 Related Work
	3 Algorithms Overview
		3.1 Convolutional Neural Network (CNN)
		3.2 Multilayer Perceptron (MLP)
		3.3 Support Vector Machine (SVM)
	4 Implementation
		4.1 Package Import
		4.2 Data Preprocessing and Handling
		4.3 Model Creation
		4.4 Model Compilation, Evaluation and Optimization
		4.5 Testing and Saving Model
		4.6 Graphical User Interface (GUI)
	5 Results
		5.1 CNN
		5.2 MLP
		5.3 SVM
		5.4 Real Time Detection from GUI
	6 Performance Analysis
	7 Conclusion
	References
Towards Automatic Assessment of Quiet Standing Balance During the Execution of ADLs
	1 Introduction
	2 Related Work
	3 CoM Estimation
		3.1 System Proposal
		3.2 Experimental Setup
		3.3 Results
	4 Score Estimation of the Quiet Standing Postural Stability
		4.1 System Proposal
		4.2 Experimental Setup
		4.3 Results
	5 Conclusions and Future Work
	References
Deep/Machine Learning in Engineering
Adaptive Model for Industrial Systems Using Echo State Networks
	1 Introduction
	2 Echo State Networks
	3 Methodology
		3.1 Quadruple-Tank Process
		3.2 Adaptive Model
	4 Results
	5 Conclusions
	References
DNN-Driven Gradient-Based Shape Optimization in Fluid Mechanics*-12pt
	1 Introduction
	2 Methods and Tools
		2.1 CFD and Shape Parameterization Tools
		2.2 DNNs - Training and Differentiation
		2.3 The Proposed DNN-Driven Gradient-Based Algorithm
	3 Applications in Aerodynamic Shape Optimization
		3.1 Problem I: Inviscid Flow Around an Airfoil
		3.2 Problem II: Laminar Flow Within an S-Bend Duct
	4 Conclusions
	References
Residual Error Learning for Electricity Demand Forecasting
	1 Introduction
	2 Relevant Work
	3 Proposed Method
	4 Experimental Evaluation
		4.1 Datasets
		4.2 Models
		4.3 Evaluation Metrics
		4.4 Experimental Setup
		4.5 Implementation Details
		4.6 Experimental Results
	5 Conclusions
	References
Strain Prediction of a Bridge Deploying Autoregressive Models with ARIMA and Machine Learning Algorithms
	1 Introduction
	2 Area of Research
	3 Dataset
		3.1 Dataset Preprocessing
	4 Algorithms and Evaluation Indices
		4.1 Machine Learning Algorithms and ARIMA Description
		4.2 Evaluation of Deep Learning Algorithms.
	5 Evaluation and Experimental Results
	6 Conclusions and Future Work
	References
Verification of Neural Networks Meets PLC Code: An LHC Cooling Tower Control System at CERN
	1 Introduction
	2 Background
		2.1 Verification of NNs
		2.2 PLCverif
	3 Case Study: The LHC Cooling Towers Controls
		3.1 Control Design for the Cooling Towers
		3.2 Approximate MPC Using Neural Networks
	4 Verification of a NN-Based Controller on a PLC
		4.1 Properties to Verify
		4.2 Verification of a NN with PLCverif
		4.3 Verification of a NN Using Other Methods
	5 Empirical Results
		5.1 Comparison of the Different Approaches
	6 Conclusions and Future Work
	References
Virtual Flow Meter for an Industrial Process
	1 Introduction
	2 Literature Review
	3 Methodology
		3.1 The Industrial System: A Pilot Plant
		3.2 Measurement of Flow and Data Acquisition
		3.3 Developing a Virtual Flow Meter
	4 Experimental Results and Discussion
	5 Conclusions
	References
Wind Energy Prediction Guided by Multiple-Location Weather Forecasts
	1 Introduction
	2 Proposed Method
		2.1 Problem Statement and Notations
		2.2 Multi-Kernel Convolutional Scaled Dot-Product Attention
		2.3 Model Architecture
	3 Experimental Evaluation
		3.1 Dataset Description
		3.2 Baseline Methods
		3.3 Experimental Results
	4 Conclusions
	References
Learning (Reinforcemet - Federated - Adversarial - Transfer)
An Autonomous Self-learning and Self-adversarial Training Neural Architecture for Intelligent and Resilient Cyber Security Systems
	1 Introduction
	2 Methodology
	3 Dataset and Results
	4 Conclusion
	Appendix 1
	References
Deep Transfer Learning Application for Intelligent Marine Debris Detection
	1 Introduction
	2 Methodology
		2.1 Image Dataset
	3 Transfer Learning of Models
	4 Comparison of Models
	5 Deployment
	6 Conclusion
	References
Forecasting Functional Time Series Using Federated Learning
	1 Introduction
	2 Literature Review
		2.1 Functional Time Series Forecasting
		2.2 Federated Learning
	3 Forecasting Functional Time Series Using Federated Learning
		3.1 Functional Neurons
		3.2 Functional Multilayer Perceptron
		3.3 Federated Averaging: FedAvg
	4 Experimental Setup
		4.1 Remaining Useful Life Estimation of Aircraft Engines
		4.2 Horizontal Data Partitioning
	5 Results and Analysis
		5.1 Centralized Data Scenario
		5.2 Federated Scenarios
	6 Conclusion and Future Works
	References
Group-Personalized Federated Learning for Human Activity Recognition Through Cluster Eccentricity Analysis*-12pt
	1 Introduction
	2 Related Work
	3 Preliminaries
		3.1 Baselines
		3.2 Problem Setting
		3.3 Data Smoothing
		3.4 Markov Clustering Algorithm
		3.5 Wasserstein Distance
		3.6 Eccentricity Analysis
	4 Proposed Approach
	5 Experimental Design
		5.1 HAR Datasets
		5.2 Evaluation Strategy
	6 Experimental Results
	7 Conclusion
	References
Modeling Others as a Player in Non-cooperative Game for Multi-agent Coordination
	1 Introduction
	2 Related Work
	3 Background
		3.1 Markov Game and MARL
		3.2 Bi-level Actor-Critic
		3.3 Two-Player Non-cooperative Game with MOA
	4 Proposed Method
		4.1 MOA with Historical Information
		4.2 Markov Game with MOA
	5 Experiments
		5.1 Experimental Environment
		5.2 Experimental Results
	6 Conclusion
	References
Neural Network Bootstrap Forecast Distributions with Extreme Learning Machines*-12pt
	1 Introduction
	2 Bootstrap Forecast Densities with Neural Networks
	3 Extreme Learning Machines
	4 Computational Burden and Algorithm Scalability
	5 Simulation Results
	6 Concluding Remarks
	References
Subsampled Dataset Challenges and Machine Learning Techniques in Table Tennis
	1 Introduction
	2 Methodology
		2.1 Dataset Representation
		2.2 Dataset Analysis
		2.3 Implemented Machine Learning Algorithms
	3 Implementation
		3.1 Model Structures
		3.2 Results Comparison
	4 Conclusion
	References
Towards Explaining Shortcut Learning Through Attention Visualization and Adversarial Attacks*-12pt
	1 Introduction
	2 Background
		2.1 Attention Mechanism
		2.2 Shortcut Learning
		2.3 Visualizing Attention Weights
	3 AttentiveBERT
		3.1 Tasks
		3.2 Models and Configuration
		3.3 Single Input Visualization
		3.4 Compare Inputs Visualization
	4 Generating Adversarial Attacks
	5 Conclusion
	References
Natural Language - Recommendation Systems
A Novel Neural Network-Based Recommender System for Drug Recommendation
	1 Introduction
	2 Related Work
	3 Designing the Recommender System
		3.1 Loading and Preprocessing
		3.2 Model Selection and Training
		3.3 Recommendation Score Determination
	4 Results and Evaluation
		4.1 Evaluation
		4.2 Comparison with an Existing Model
	5 Conclusion
	References
DACL: A Domain-Adapted Contrastive Learning Approach to Low Resource Language Representations for Document Clustering Tasks
	1 Introduction
	2 Related Work
	3 Materials and Methods
		3.1 Datasets
		3.2 Domain-Adapted Contrastive Representation Learning
		3.3 Generating Representations
	4 Results and Discussion
		4.1 Experimental Setup
		4.2 Experimental Results
	5 Conclusions
	References
Evaluating the Extraction of Toxicological Properties with Extractive Question Answering
	1 Introduction
	2 Related Works
	3 Question Answering for Information Extraction
	4 Experimentation and Results
		4.1 Setup
		4.2 Results and Discussion
	5 Conclusion
	References
Semi-automated System for Pothole Detection and Location
	1 Introduction
		1.1 Objective
		1.2 State of Art
	2 Methodology
		2.1 AI Model
		2.2 Filtration Program
		2.3 Obtaining Geographical Coordinates
		2.4 Point Mapping
	3 Results and Analysis
		3.1 AI Model
		3.2 Data Generation
		3.3 Final Maps
	4 Conclusions
	References
Author Index