Advances in Computational Collective Intelligence : 14th International Conference, ICCCI 2022, Hammamet, Tunisia, September 28–30, 2022, Proceedings

This book constitutes refereed proceedings of the 14th International Conference on International Conference on Computational Collective Intelligence, ICCCI 2022, held in Hammamet, Tunisia, in September 2022. The 43 full papers and 15 short papers were thoroughly reviewed and selected from 421 submissions. The papers are grouped in topical ​sections on ​collective intelligence and collective decision-making; natural language processing; deep learning; computational intelligence for multimedia understanding; computational intelligence in medical applications; applications for industry 4.0; experience enhanced intelligence to IoT and sensors; cooperative strategies for decision making and optimization; machine learning methods. Preface Organization Contents Collective Intelligence and Collective Decision-Making From Unhealthy Online Conversation to Political Violence: The Case of the January 6th Events at the Capitol 1 Introduction 2 From Toxicity and Incivility to Unhealthy Online Conversation 3 Data and Methods 3.1 Data 3.2 Representation 3.3 Classifier Evaluation 3.4 Topic Discovery 4 Results 4.1 Attribute Distribution and Trend Analysis 4.2 Topic Analysis 4.3 Discussion 5 Conclusions References Unraveling COVID-19 Misinformation with Latent Dirichlet Allocation and CatBoost 1 Introduction 2 Preliminaries and Related Work 2.1 COVID-19 Misinformation Detection Literature 2.2 Categorical Boosting (CatBoost) 2.3 Term Frequency - Inverse Document Frequency (TF-IDF) 2.4 Latent Dirichlet Allocation (LDA) 2.5 Shapley Additive Explanations (SHAP) 3 Contribution 4 Methodology 4.1 Dataset Description and Preprocessing 4.2 Topic Modeling Using LDA 4.3 Model Training 5 Results and Discussion 5.1 Performance 5.2 Interpretability Using SHAP 6 Conclusion and Future Work References Augmentation-Based Ensemble Learning for Stance and Fake News Detection 1 Introduction 2 Text as Vectors 2.1 Pre-processing and Feature Extraction 2.2 Dimensionality Reduction 3 Text Data Augmentation 3.1 Masked Language Models 3.2 Back-translation (a.k.a. Round-Trip Translation) 3.3 Synonym (a.k.a. Thesaurus-Based Augmentation) 3.4 TF-IDF Based Insertion and Substitution 4 Augmentation-Based Ensemble Learning 4.1 Diversity and Skillfulness in Ensemble Learning 4.2 Novel Augmentation Based Approach 5 Related Work 6 Experimental Study 6.1 Tools and Datasets 6.2 Results and Discussion 7 Conclusion References An Opinion Analysis Method Based on Disambiguation to Improve a Recommendation System 1 Introduction 2 Problematic 3 Recommender Systems 3.1 Recommender System Based on Collaborative Filtering 3.2 Recommender System Based on Content Filtering 3.3 Hybrid Recommender System 3.4 Hybrid Recommender System 3.5 State of the Art on Recommender Systems 4 Proposed Method 4.1 Proposed Method Algorithm 4.2 Description of the Proposed Method 4.3 Collaborative Filtering 4.4 Automatic Language Processing 5 Experimentation, Results and Discussion 5.1 Corpus Description 5.2 Evaluation of the Referral System 6 Conclusion 7 Limits and Perspectives References A Recommendation System for Job Providers Using a Big Data Approach 1 Introduction 2 Related Work 3 Proposed Model 4 Experiments 4.1 Datasets 4.2 Experimental Results 5 Conclusion References Natural Language Processing Multi-module Natural Language Search Engine for Travel Offers 1 Introduction 2 Related Work 3 Metadata Module 3.1 Datasets 3.2 Method 3.3 Performance 4 Ontology-Based Text Matching Module 4.1 Ontology and Datasets 4.2 Method 4.3 Performance 5 Semantic Vector Search Module 5.1 Datasets 5.2 Method 5.3 Performance 6 Combination of Modules 6.1 Combination Architecture 6.2 Performance 7 Conclusions and Future Work References Towards Automatic Detection of Inappropriate Content in Multi-dialectic Arabic Text 1 Introduction 2 Related Work 3 Tun-EL Dataset Construction 3.1 Dataset Selection 3.2 Dataset Processing 3.3 Exploratory Data Analysis 4 Inappropriate Multidialect Content Detection 4.1 Mono-dialect vs Multi-dialect Training Approaches 4.2 Proposed Model 5 Experiments and Evaluation 5.1 Data Preprocessing 5.2 Experimental Setup 5.3 Results and Discussion 6 Conclusion and Perspectives References Improving Bert-Based Model for Medical Text Classification with an Optimization Algorithm 1 Introduction 2 Related Work 3 Proposal Approach 3.1 Pre-processing 3.2 Bert-Based Model for Textual Feature Extraction 3.3 Medical Text Classification 3.4 Hyperparameter Selection 4 Experiment Results and Discussion 4.1 Data Description and Evaluation Metrics 4.2 Evaluation of the Classification Method 4.3 Evaluation of the Hyperparameter Selection Method 5 Conclusion References Reinforcement of BERT with Dependency-Parsing Based Attention Mask 1 Introduction 2 Transformers 2.1 Scaled Dot-Product Attention Mechanism 2.2 Padding Mask 3 Proposed Mask 4 Experimentations 5 Conclusion References BERT and Word Embedding for Interest Mining of Instagram Users 1 Introduction 2 Related Work 3 Dataset Overview 3.1 Collecting Data 3.2 Data Description 3.3 Labelling Data 3.4 Data Analysis 4 Task Modeling 4.1 Instagram Model Formalization 4.2 Task Formulation 5 Extracting User Interests 5.1 Feature-Based Variant 5.2 Bert-based Variant 6 Experiments 6.1 Data Pre-processing and Implementation 6.2 Results 7 Conclusion and Future Work References Multi-Wiki90k: Multilingual Benchmark Dataset for Paragraph Segmentation 1 Introduction 2 Related Works 3 Multilingual Benchmark Dataset 3.1 Existing Datasets 3.2 Multi-Wiki90k 4 Models 4.1 LASER 4.2 LaBSE 4.3 mBERT 4.4 XLM-RoBERTa 5 Evaluation 5.1 Configuration 5.2 Results 6 Conclusions and Future Work References CAE: Mechanism to Diminish the Class Imbalanced in SLU Slot Filling Task 1 Introduction 2 Related Work 2.1 SLU Task 2.2 Class Imbalance in Sequence Labeling 3 Methodology 3.1 Baseline Model 3.2 Oversampled Data 3.3 Proposed Model 4 Experiments and Analysis 4.1 Experimental Settings 4.2 Experimental Results 4.3 Analysis 5 Conclusion References Deep Learning An Image Retrieval System Using Deep Learning to Extract High-Level Features 1 Introduction 2 Related Work 3 Proposed Method 3.1 Transfer Learning 3.2 Convolutional Neural Networks 4 Experiments 4.1 Dataset 4.2 Image Descriptor 4.3 Indexing 4.4 Similarity Measure 5 Perfomance Measure 6 Conclusion References An Effective Detection and Classification Approach for DoS Attacks in Wireless Sensor Networks Using Deep Transfer Learning Models and Majority Voting 1 Introduction 2 Related Work 3 Dataset 4 Proposed Approach 4.1 Data Preprocessing 4.2 Data Transformation 4.3 Transfer Learning 4.4 Ensemble Classifier Based on Majority Voting 5 Experiments 5.1 Implementation Details 5.2 Performance Metrics 5.3 Results 5.4 Comparison with Existing ML/DL-based Approaches 6 Conclusion References SSTop3: Sole-Top-Three and Sum-Top-Three Class Prediction Ensemble Method Using Deep Learning Classification Models 1 Introduction 2 Related Works 3 Proposed Methodology 4 Experiments and Results 4.1 Evaluation Metrics 5 Conclusion References Hyperparameter Optimization of Deep Learning Models for EEG-Based Vigilance Detection 1 Introduction 2 Materials and Methods 2.1 EEG Signal Acquisition and Pre-processing 2.2 DL Models Hyperparameters 2.3 HPO Algorithms 3 Experiments and Results 3.1 Experiment Setting 3.2 Results and Discussion 4 Conclusion and Perspectives References A Hybrid Face Recognition Approach Using Local Appearance and Deep Models 1 Introduction 2 Methodology 2.1 Face Detection 2.2 Face Alignment 2.3 Feature Extraction 2.4 Distance Metric Fusion 3 Experiments 3.1 Dataset Description 3.2 Implementation Details 3.3 Experimental Results 4 Conclusion References Deep Learning-Based Text Recognition of Agricultural Regulatory Document 1 Introduction 2 Related Studies 2.1 Text Detection and Recognition 3 Methodology 3.1 Image Pre-processing 3.2 Evaluation of Text Detection and Recognition Models 4 Models 4.1 Character Region Awareness for Text (CRAFT) 4.2 E2E-MLT OCR Branch 4.3 Sliding Windows Text Matcher 5 Results 6 Real World Deployment 7 Conclusion References Computational Intelligence for Multimedia Understanding Textural Features Sensitivity to Scale and Illumination Variations 1 Introduction 2 Markovian Textural Features 3 Frequented Alternative Features 4 Experiments 4.1 University of East Anglia Uncalibrated Image Database 4.2 Wood UTIA BTF Database 4.3 Setup 4.4 Results 5 Conclusion References Recognizing Handwritten Text Lines in Ancient Document Images Based on a Gated Residual Recurrent Neural Network 1 Introduction 2 Related Work 3 Proposed Framework 4 Experiments and Results 4.1 Experimental Corpora 4.2 Experimental Protocol 4.3 Evaluation Metrics 4.4 Results 5 Conclusions and Further Work References Damage Detection of Coated Milling Tools Using Images Captured by Cylindrical Shaped Enclosure Measurement Setup 1 Introduction 2 Coated Milling Tools 2.1 Tool Specifications 3 Measurement Setup 4 Algorithm for Damage Detection 5 Results 6 Conclusion References Efficient Machine-Learning Based 3D Face Identification System Under Large Pose Variation 1 Introduction 2 Related Works 3 Pose Normalization 4 Description of the Proposed System 5 Performance of the Proposed System 6 Conclusion References Arabic Handwritten Character Recognition Based on Convolution Neural Networks 1 Introduction 2 Related Work 3 Proposed CNN Model 3.1 Input Layer 3.2 Hidden Layers 3.3 Output Layer 4 Experimental Results 4.1 Hijja Dataset 4.2 4.2. Results and Discussion 5 Conclusion References An End to End Bilingual TTS System for Fongbe and Yoruba 1 Introduction 2 Objectives 3 Related Work 4 Model Architecture 5 Datasets 5.1 Fongbe Language 5.2 Yoruba Language 5.3 Characteristics of the Datasets 6 Experiments 6.1 Monolingual Training 6.2 Multilingual Training 6.3 Voice Cloning 7 Conclusion References Application of the Laplacian Smoothing on 3D Static Models and Their Evaluation by the New Objective Quality Metric 3DrwPSNR 1 Introduction 2 State of the Art 3 Proposed Approach 4 Experiments 5 Conclusion References Computational Intelligence in Medical Applications Brain Tumors Detection on MRI Images with K-means Clustering and Residual Networks 1 Introduction 2 Related Work 3 Methodology 3.1 The Research Implementation Procedure 3.2 K-means Clustering 3.3 Data Augmentation 3.4 The Proposed CNN Architecture 4 Experiments 4.1 Dataset 4.2 Comparison with Some Well-Known CNN Architectures 5 Conclusion References Approximating Sparse Semi-nonnegative Matrix Factorization for X-Ray Covid-19 Image Classification 1 Introduction 2 Sparse Semi-NMF as a Linear Optimisation Problem 3 The Algorithm 4 Experimental Results 5 Conclusion References Hybrid Architecture for 3D Brain Tumor Image Segmentation Based on Graph Neural Network Pooling 1 Introduction 2 Related Work 2.1 Classical Segmentation Methods 2.2 Deep Learning Segmentation Methods 3 Proposed Method 3.1 Feature Extraction via 3DBUNet 3.2 Graph Module 4 Pre-processing and Image Segmentation Results 4.1 Datasets Pre-processing: BraTS Annotations and Structures 4.2 Experimental Results 5 Conclusion References Right Ventricle Segmentation in Cardiac MR Images Using Convolutional Neural Network Architecture 1 Introduction 2 State of the Art 3 Methodology 3.1 DataSet 3.2 Preprocessing 3.3 Data Augmentation 3.4 Inter-Expert Similarity 3.5 Algorithm 4 Result 5 Conclusion References Fast Unsupervised Residual Attention GAN for COVID-19 Detection 1 Introduction 2 Proposed Method 2.1 Model Overview 2.2 Attention Mechanism 2.3 Model Architecture 2.4 Model Training 2.5 Anomaly Scores 3 Datasets and Experimental Setup 3.1 Datasets 3.2 Experimental Setup 4 Experimental Results 5 Conclusion References Detection of Breast Masses in Mammograms by Incremental Discriminant Based Support Vector Machine Classifier and Active User Involvement 1 Introduction 2 Related Work 3 Mass Detection by IDSVM 3.1 Mammographic Images Preprocessing 3.2 Feature Engineering 3.3 IDSVM Based Classification 4 Results and Discussion 4.1 Masses Shape Influence 4.2 Masses Size Influence 4.3 Breast Tissue Density Influence 5 Conclusion References Analysis of Right Ventricle Segmentation in the End Diastolic and End Systolic Cardiac Phases Using UNet-Based Models 1 Introduction 2 Related Works 3 Proposed Method 3.1 Datasets Description 3.2 Architecture 4 Experimental Results and Discussion 4.1 Evaluation Metrics 4.2 Quantitative Evaluation 4.3 Qualitative Evaluation 4.4 Comparison with State of the Art Methods 5 Conclusion References An Original Continuous-to-Continuous Forward Model as a Universal Method for the Formulation of Reconstruction Methods for Medical Imaging Techniques 1 Introduction 2 Continuous-to-Continuous Data Model Formulation 3 Model-based Iterative Reconstruction Method 3.1 Model-based Iterative Reconstruction Method for PET 3.2 Model-based Iterative Reconstruction Method for CT 4 Experimental Results 5 Conclusions References Applications for Industry 4.0 Towards a Dynamic Vehicular Clustering Improving VoD Services on Vehicular Ad Hoc Networks 1 Introduction 2 Related Work 3 The Proposed Solution 3.1 System Model and Problem Formulation 3.2 Dynamic Vehicles Clustering Features 4 Simulation Results and Analysis 5 Conclusion References RVT-Transformer: Residual Attention in Answerability Prediction on Visual Question Answering for Blind People 1 Introduction 2 Related Works 3 Our Model 3.1 Feature Extraction 3.2 Residual Attention 3.3 Residual VT-Transformer 4 Experiments 4.1 Dataset and Evaluation Metrics 4.2 Experimental Settings 4.3 Results 4.4 Ablation Studies 5 Discussion 6 Conclusion References Automatic Processing of Planning Problems: Application on Representative Case Studies 1 Introduction 2 Related Works 3 Automatic Processing of Planning Problems 3.1 Domain and Problem Constructs 3.2 Planner and Validator 4 Formal Planning of Hanoï Towers 4.1 Planning Domain Hanoï 4.2 The Three-Disk Hanoï Problem 4.3 Plan-Solution 5 Formal Planning of Sliding Puzzle Game 5.1 First Modeling 5.2 Second Modeling 5.3 Comparison 6 Conclusion References Application of Software Defined Networks for Collection of Process Data in Industrial Real-Time Systems 1 Introduction 2 Background 3 Software Defined Networks 4 Process Data Extraction Using SDN 5 Testbed 6 Measurements and Results 7 Conclusions References ITS Traffic Violation Regulation Based on Blockchain Smart Contracts 1 Introduction 2 Related Work 3 Smart Contract-Based Regulation of Traffic Violations in ITS 4 Experiment and Discussion 5 Conclusion and Perspective References Distributed Architecture of an Intrusion Detection System in Industrial Control Systems 1 Introduction 2 Related Work 3 A New Distributed Approach of Intrusion Detection System in Industrial Control Systems 3.1 Industrial Data Collection 3.2 Data Storage 3.3 Data Structuring and Preprocessing 3.4 Intrusion Detection and Classification 4 Evaluation and Validation of Results 4.1 Evaluation Metrics 4.2 Results 4.3 Comparison and Discussion of Results 5 Conclusion and Future Work References MAFC: Multimedia Application Flow Controller for Big Data Systems 1 Introduction 2 System Architecture and Operations 2.1 Receiver 2.2 Flow Controller 2.3 Policy Adapter 2.4 Control Policies Database 3 Application to Surveillance Systems 3.1 MAFC Operations and Flow Control Policies in a Surveillance System 3.2 Implementation 4 Experimental Evaluation 4.1 Processing Time Evaluation 4.2 Infrastructure Resources' Impact 4.3 Flow Control Scalability 5 Related Works 6 Conclusion References Experience Enhanced Intelligence to IoT and Sensors On-wrist Based Datasets Exploration for an IoT Wearable Fall Detection 1 Introduction 2 Methodology: IoT-Based Framework for Wrist-Based Fall Detection System Using SDL 2.1 Overview of the Proposed Framework 2.2 Dataset Exploration 2.3 Data Preprocessing 2.4 Dictionary Learning-Based Classification for Wristband Fall Detection 3 Experimental Test and Evaluation 3.1 Evaluation Metrics 3.2 Experimental Setup 3.3 Offline Experiment 3.4 Prototype 4 Conclusion References UML Profile for IoT-Based Applications 1 Introduction 2 Related Work 3 Background 3.1 UML Profile 3.2 IoT 4 Proposed IoT System Profile 5 OCL IoTSystem Profile Constraints Definition 5.1 «IoTSystem »stereotype 5.2 «IoTComponent »stereotype 5.3 «IoTPort »stereotype 5.4 «IoTInterface »stereotype 6 Use Case: A Smart Home System 7 Conclusion References A Formal Verification Model for IoT Based Applications Using Event-B 1 Introduction 2 Related Work 3 Event-B Method 4 Informal Description 4.1 Structural Requirements 4.2 Behavioral Requirements 5 Event-B Formal Model 5.1 Formal Development 5.2 Verification 6 Conclusion and Future Work References Machine Learning and IoT for Stress Detection and Monitoring 1 Introduction 2 Related Work 3 Materials and Methods 3.1 Data Collection 3.2 Features Extraction 3.3 Analysis Methods 3.4 Models Evaluation and Performance Measures 4 Design and Implementation 4.1 Acquisition Layer 4.2 Service Layer 4.3 AI Layer 5 Results and Discussion 6 Conclusion and Future Work References Long Short-Term Memory Based Photoplethysmography Biometric Authentication 1 Introduction 2 Literature Review 3 Methodology 3.1 Signal Preprocessing 3.2 PPG Data Preparation 3.3 Proposed Deep Learning Solution for Feature Extraction and Classification: 4 Experimental Results 5 Conclusion References Extended U-net for Retinal Vessel Segmentation 1 Introduction 2 Related Works 3 Patch Extraction 4 Proposed Architecture for Retinal Blood Vessel Segmentation 4.1 Convolution Processing 4.2 Proposed Network 4.3 Training Parameter Setting 5 Experiments 5.1 Dataset, Evaluation Metrics and Experiment Setup 5.2 Identification of Convolution Kernel Size 5.3 Segmentation Performance 5.4 Comparison with State of the Art 6 Conclusion References Cooperative Strategies for Decision Making and Optimization Spatial Clustering by Schelling's Ants 1 Introduction 2 Spatial Clustering by Ants 3 Methods 4 Results 4.1 Dependence on Parameters 5 Conclusion and Future Works References A New Variant of the Distributed Permutation Flow Shop Scheduling Problem with Worker Flexibility 1 Introduction 2 Related Literature Review 2.1 Distributed Permutation Flow Shop Scheduling Problem 2.2 Flow Shop Scheduling Problem with Worker Flexibility 3 Description of the Distributed Permutation Flow Shop Scheduling Problem with Worker Flexibility 4 Heuristic Method for the DPFSPw 5 Conclusion and Future Directions References Resource Allocation Strategy on Yarn Using Modified AHP Multi-criteria Method for Various Jobs Performed on a Heterogeneous Hadoop Cluster 1 Introduction 2 Related Work 3 Dynamic Multi-criteria Decision-Making for Resource Allocation in a Heterogeneous Yarn Cluster 4 Problem Statement 5 Proposed Algorithm 6 Experimental Results 7 Conclusion and Future Works References A Memetic Approach for Routing Problem with Capacity and Time Constraints 1 Introduction 2 Problem Description 3 Greedy Randomized Adaptive Search Procedure 4 Chemical Reaction Optimization 4.1 Basic Concepts 4.2 CRO Algorithm 5 Memetic Approach 5.1 Solution Representation 5.2 Initialization Stage 5.3 Optimization Process 5.4 Collision Operators 6 Experimental Study 7 Conclusions References A Novel Unfeasible Space Exploring Matheuristic Proposal to Solve the Sum Coloring Problem 1 Introduction 1.1 Review and State-of-the-art 1.2 Contributions 2 Problem Statement: Preliminary Definitions and Properties 3 A Novel Weight-Based Formulation WBF 4 Methodology 4.1 The Proposed Matheuristic Approach 4.2 An Iterated Simulated Annealing for the Sum Coloring Problem: ISA 5 Computational Study 5.1 Experimental Setup and Tested Instances 5.2 Used Metrics 5.3 Evaluation of the Proposed Matheuristic and Comparisons 6 Conclusion References Machine Learning Methods Addressing the Complexity of MOBA Games Through Simple Geometric Clues 1 Introduction 2 Related Work 3 Addressing the Complexity Through Geometric Clues 3.1 WrapperApproach with RNN 3.2 Experimental Conditions 4 Results 4.1 Learning Since the Beginning 4.2 Sliding Window Learning 5 Conclusion References Bots and Gender Detection on Twitter Using Stylistic Features 1 Introduction 2 Related Work 2.1 Bots Detection 2.2 Gender Identification 3 Proposed Approach 3.1 Pre-processing 3.2 Features Extraction 4 Experimental Study and Result Analysis 4.1 Datasets 4.2 Classification Algorithms and Hyperparameters Tuning 4.3 Results Analysis 4.4 Comparative Study 5 Conclusion References How Differential Privacy Reinforces Privacy of Machine Learning Models? 1 Introduction 2 Membership Inference Attacks (MIA) 2.1 Overview on MIAs 2.2 Factors Influencing the Risk of Membership Inference Attacks 3 Differential Privacy 3.1 Overview on DP 3.2 Differentially Private SGD Algorithm 4 Related Works 5 Our Contribution 5.1 Dataset 5.2 Implementation 5.3 Evaluation of Our Strategy Against MIAs and Discussion 6 Conclusion References Stochastic Expectation Propagation Learning for Unsupervised Feature Selection 1 Introduction 2 The Infinite GID Mixture with Feature Selection 2.1 Infinite GID Mixture Models with Feature Selection 2.2 Prior Distributions 3 Model Learning via SEP 4 Experimental Results 4.1 Simulated Data 4.2 Real Data 4.3 Occupancy Detection and Estimation in Smart Buildings 5 Conclusion References S-LDA: Documents Classification Enrichment for Information Retrieval 1 Introduction 2 Related Works 3 Methodology 3.1 Document Structure Analysis 3.2 Document Classification 4 Experiments 4.1 Environment 4.2 Evaluation Metrics 4.3 Experimental Protocol 5 Conclusion References Distributed Anomalies Detection Using Isolation Forest and Spark 1 Introduction 2 Isolation Forest Algorithm 2.1 Training Phase 2.2 Scoring Phase 3 Distributed Isolation Forest 3.1 Spark 3.2 Description of LDIForest and FDIForest 3.3 Experimental Comparison Between LDIForest and FDIForest 4 Conclusion and Future Work References Defining and Extracting Singleton Design Pattern Information from Object-Oriented Software Program 1 Introduction 2 Related Work 3 Proposed Features for Singleton Implementation Variants 3.1 Singleton Variants 3.2 Proposed Features 4 RNN-LSTM Classifier for Structural and Semantic Source Code Analysis 4.1 Program Analysis 4.2 RNN-LSTM Classifier for Features Extraction 5 Experimental Setup 5.1 Evaluation Protocol 5.2 Experimental Results 6 Conclusion and Future Works References Author Index