Intelligent Information and Database Systems: 15th Asian Conference, ACIIDS 2023, Phuket, Thailand, July 24–26, 2023, Proceedings, Part I (Lecture Notes in Artificial Intelligence)

This two-volume set LNAI 13995 and LNAI 13996 constitutes the refereed proceedings of the 15th Asian Conference on Intelligent Information and Database Systems, ACIIDS 2023, held in Phuket, Thailand, during July 24–26, 2023. The 65 full papers presented in these proceedings were carefully reviewed and selected from 224 submissions. The papers of the 2 volume-set are organized in the following topical sections: Case-Based Reasoning and Machine Comprehension; Computer Vision; Data Mining and Machine Learning; Knowledge Integration and Analysis; Speech and Text Processing; and Resource Management and Optimization. Preface Organization Contents – Part I Contents – Part II Case-Based Reasoning and Machine Comprehension On the Improvement of the Reasoning Cycle in Case-Based Reasoning 1 Introduction 2 A More Domain Knowledge Independent Approach 2.1 Case Structure 2.2 Retrieving Similar Cases 2.3 An Optimization Approach for Adaptation 3 Experimental Evaluation 3.1 Case Study 3.2 Dataset Description 3.3 Testbed Setup 3.4 Empirical Evaluation 4 Conclusion References Exploring Incompleteness in Case-Based Reasoning: A Strategy for Overcoming Challenge 1 Introduction 2 Background 2.1 Case-Based Reasoning and Data Completeness 2.2 Change Point Analysis 3 The Completeness Challenge: A Problem Statement 3.1 Illustrative Scenario 3.2 Problem Formulation 4 Evaluating Data Incompleteness in CBR Systems 4.1 Segmentation of the Case Base 4.2 Identification of Incompleteness Situations 5 Experimental Results 6 Conclusion References Leveraging both Successes and Failures in Case-Based Reasoning for Optimal Solutions 1 Introduction 2 Illustrative Example and Preliminary Concepts 2.1 Key Concepts and Notations Related to the CBR Paradigm 2.2 Collision Avoidance Navigation 3 Adaptation Through Failed and Successful Cases 3.1 Problem Statement 3.2 Principle 3.3 Local Prediction of the Target Solution 4 Evaluation 4.1 Experimental Design 4.2 Baselines and Metrics 4.3 Results 5 Conclusion References Transfer Learning for Abnormal Behaviors Identification in Examination Room from Surveillance Videos: A Case Study in Vietnam*-1pc 1 Introduction 2 Related Work 3 Methods 3.1 Abnormal Behaviors in the Examination Hall and Data Collection 3.2 Deep Learning Architectures for Abnormal Behavior Detection 4 Experimental Results 4.1 Environmental Settings 4.2 Abnormal Behavior Detection Using YOLO V4 4.3 Abnormal Behavior Detection with SSD MobileNet V2 4.4 Discussion 5 Conclusion References A Novel Question-Context Interaction Method for Machine Reading Comprehension 1 Introduction 2 Related Work 3 Methodology 3.1 Sentence Embedding 3.2 S-QCI Block 3.3 Word Fusion: 3.4 Training Process 4 Experiment 4.1 Experimental Setup 4.2 Benchmark Dataset and Baseline Models 4.3 Main Results 4.4 Ablation Study 5 Conclusion References Granular Computing to Forecast Alzheimer’s Disease Distinctive Individual Development 1 Introduction 2 Methods 2.1 Rough Set Implementation of GrC 3 Results 3.1 Statistics 3.2 Granular Computing for Reference of Group1 Group 3.3 Granular Computing for Reference of Group2 Patients 4 Discussion References Computer Vision AdVLO: Region Selection via Attention-Driven for Visual LiDAR Odometry 1 Introduction 2 Related Works 2.1 Visual Odometry 2.2 LiDAR Odometry 2.3 Visual LiDAR Odometry 3 Multimodel SLAM with Attention 3.1 Problem Description 3.2 Feature Embedding 3.3 Attention-Driven Region Selection 3.4 Loss Function 4 Experiments 4.1 Experiments Setting 4.2 Experimental Results 5 Conclusion References Intelligent Retrieval System on Legal Information 1 Introduction 2 Related Work 3 Methodology 3.1 Problem Formulation 3.2 Data Collection 3.3 Data Processing 3.4 Vocabulary Frequency 3.5 Feature Extraction and Modeling 3.6 Metric Evaluation 4 Experiments 4.1 Experimental Design 4.2 Results 5 Conclusion and Future Work References VSNet: Vehicle State Classification for Drone Image with Mosaic Augmentation and Soft-Label Assignment 1 Introduction 2 Related Work 2.1 Autonomous Vehicle Dataset for Object Detection 2.2 Drone-Based Dataset for Object Detection 3 Proposed Algorithm 3.1 Vehicle Detection 3.2 Mosaic Data Augmentation 3.3 Soft Label Assignment 3.4 Vehicle State Classification 4 Experiment 5 Conclusion References Creating High-Resolution Adversarial Images Against Convolutional Neural Networks with the Noise Blowing-Up Method 1 Introduction 1.1 Standard Methodology 1.2 Three Challenges 1.3 Our Contribution 1.4 Organisation of the Paper 2 CNNs and Attack Scenarios 2.1 Assessment of the Human Perception of Distinct Images 2.2 Attack Scenarios in the R Domain 2.3 Attack Scenarios Expressed in the H Domain 3 The Noise Blowing-Up Strategy 3.1 Construction of Adversarial Images in H for the Target Scenario 3.2 Indicators 4 Case Study 4.1 The CNN, the Scenario, the Images 4.2 The Attack 4.3 Experimental Results 5 One Detailed Example 6 Conclusion References Faster Imputation Using Singular Value Decomposition for Sparse Data 1 Introduction 2 Related Works 3 Methodology 3.1 Sparse Data 3.2 Mechanisms of Missing Data 3.3 Singular Value Decomposition (SVD) 3.4 SVD Imputation (SVDI) 4 Experiments 4.1 Datasets 4.2 Experimental Design 4.3 Results and Discussion 5 Conclusion and Future Works References Combination of Deep Learning and Ambiguity Rejection for Improving Image-Based Disease Diagnosis 1 Introduction 2 Related Works 3 Proposed Methodology 3.1 Overview Approach 3.2 Feature Extraction and Classification 3.3 Imbalanced Data Processing 3.4 Ambiguity Rejection 4 Experimental Results and Analysis 4.1 Materials and Preprocessing 4.2 Evaluation Metrics 4.3 Evaluation Results and Analysis 5 Conclusions Appendix References Data Mining and Machine Learning Towards Developing an Automated Chatbot for Predicting Legal Case Outcomes: A Deep Learning Approach 1 Introduction 2 Related Work 3 Experimentation 3.1 Data Retrieval 3.2 Data Preprocessing 3.3 Semantic Analysis 3.4 Secondary Data Processing 3.5 Model Training 3.6 Model Evaluation 4 Results and Discussion 4.1 Semantic Analysis 4.2 Predictive Analysis 4.3 CNN with LDA 4.4 CAPSULES with LDA 5 Conclusions and Recommendations References Fuzzy-Based Factor Evaluation System for Momentum Overweight Trading Strategy 1 Introduction 1.1 Background and Motivation 1.2 Purpose 2 Literature Review 2.1 Development of Trading Strategy and Profitability Indicator 2.2 Fuzzy Set Theory 3 Proposed Trading Strategy Factor Valuation System 4 Experiment Results 4.1 Data Usage 4.2 Simulation Results of Random Trading and Profitability Indicator 4.3 Effectiveness of Fuzzy Quantification Module 5 Conclusions References Enhancing Abnormal-Behavior-Based Stock Trend Prediction Algorithm with Cost-Sensitive Learning Using Genetic Algorithms 1 Introduction 2 Related Work 2.1 Transfer Learning and Negative Transfer 2.2 Cost Sensitive Learning 3 Proposed Enhanced GA-Based Algorithm 3.1 Framework of the Proposed Approach 3.2 Chromosome Representation 3.3 Fitness Evaluation 3.4 Genetic Operation 4 Experimental Results and Analysis 4.1 Dataset Description and Environment Setting 4.2 Comparison Results of the Proposed and the Existing Approaches 5 Conclusion References Leveraging Natural Language Processing in Persuasive Marketing 1 Introduction 2 Literature Review on Persuasion Marketing and Personality 2.1 Automated Techniques for Recognizing Personality 3 Methodology 4 Results 5 Conclusions References Direction of the Difference Between Bayesian Model Averaging and the Best-Fit Model on Scarce-Data Low-Correlation Churn Prediction 1 Introduction and Motivation 1.1 Previous Work and Current Contribution 2 Background 2.1 A Scarce-Data Low-Correlation Customer Churn Problem 2.2 Description of the Data, and Data Cleaning 2.3 Bayesian Model Averaging: Same Prediction as Best Model 2.4 A Likelihood Function for Bayesian Model Averaging 2.5 Two Approximations to Bayesian Model Averaging 3 The Proposed Prediction Method 3.1 Feature Selection Difficulties in Scarce-Data Problems 3.2 Predict from the Direction of the Differences 4 Results: Raising a Threshold for Fewer Predictions 4.1 The Bigger the Difference, the Higher the Precision 4.2 Comparison with XGBoost, Deep Learning, and Others 5 Discussion and Conclusion References Tree-Based Unified Temporal Erasable-Itemset Mining 1 Introduction 2 Related Works 3 Problem Description 3.1 The Temporal Erasable Itemset Mining 3.2 Lower-Bound Strategy 4 Proposal Algorithm 5 Experimental Evaluation 6 Conclusion and Future Work References Design Recovery of Data Model Hidden in JSON File 1 Introduction 2 Related Works 3 Motivating Example 4 Design Recovery Method 4.1 Method Assumptions 4.2 Translation Algorithm 5 Method Evaluation 5.1 Short Description 5.2 Consistency Metrics 5.3 Method Performance 6 Summary and Further Works References Accurate Lightweight Calibration Methods for Mobile Low-Cost Particulate Matter Sensors 1 Introduction 2 Design and Implementation 2.1 Data Preprocessing 2.2 Base Model Design for MLCS Calibration 2.3 NAS Model Search Space for MLCS Calibration 2.4 NAS Methods for MLCS Calibration 3 Evaluation 3.1 Base Model 3.2 NAS Models 3.3 Evaluation of NAS-RE and NAS-RS for Particulate Matter Monitoring Platform 4 Discussion 5 Conclusion References Generating Music for Video Games with Real-Time Adaptation to Gameplay Pace 1 Introduction 2 Related Works 3 Requirements Analysis 3.1 The Model 3.2 Event Encoding 3.3 Conditioning Parameters 3.4 The Key 4 The Method Description 4.1 Test Application Game 5 Experiment Setup 6 Results 7 Conclusions 7.1 Aspects to Consider in Future Research References Detecting Sensitive Data with GANs and Fully Convolutional Networks 1 Introduction 2 Related Work 3 Training the Anonymisation Model 4 Results 5 Automatic Sensitive Data Detection System 6 Conclusions References An Unsupervised Deep Learning Framework for Anomaly Detection 1 Introduction 2 Related Work 3 Problem Definition 4 Our Proposed TCN3DPredictor Framework 4.1 3D-Causal Temporal Convolutional Network 4.2 Detection Mechanism 5 Experiments 5.1 Dataset 5.2 Baseline Methods 5.3 Evaluation Results 6 Conclusion References Extracting Top-k High Utility Patterns from Multi-level Transaction Databases 1 Introduction 2 Related Works 3 Preliminaries 3.1 Definitions 3.2 Problem Statement 4 Top-k MLHUP Mining 4.1 Determining the Initial Threshold 4.2 Pruning Candidates via EUCP 4.3 The mlTKO Algorithm 5 Evaluation Studies 6 Conclusions and Future Works References Lightweight and Efficient Privacy-Preserving Multimodal Representation Inference via Fully Homomorphic Encryption 1 Introduction 2 Related Work 2.1 Multimodal Machine Learning 2.2 Homomorphic Encrypted Neural Network 2.3 Homomorphic Representation Inference 3 Preliminaries 3.1 Fully Homomorphic Encryption (FHE) 3.2 The Levelled FHE Scheme - CKKS 3.3 Homomorphic Linear Layer 3.4 Multimodal Fusion Representation Learning 4 The Proposed Approach 4.1 System Model 4.2 Homomorphic TFN 4.3 Other Optimizations 5 Experiments 5.1 Experimental Setup 5.2 Experimental Results 6 Conclusion References Neural Machine Translation with Diversity-Enabled Translation Memory 1 Introduction 2 Related Work 2.1 Neural Machine Translation for Low-Resource Languages 2.2 Translation Memory-Augmented Neural Machine Translation 2.3 Retrieval for Translation Memory-Augmented Neural Machine Translation 3 Methodology 3.1 Overview System 3.2 Retrieval Model 3.3 Translation Model 4 Experiment 4.1 Experiment Setup 4.2 Main Results 5 Conclusion References GIFT4Rec: An Effective Side Information Fusion Technique Apply to Graph Neural Network for Cold-Start Recommendation 1 Introduction 2 Related Works 2.1 Cold-Start User Problems 2.2 Meta-learning 2.3 Graph Neural Network 3 Proposal Model 3.1 General Side Information Module 4 Experiments 4.1 Experiment Setting 4.2 Experiment Result 5 Conclusion References A Decision Support System for Improving Lung Cancer Prediction Based on ANN 1 Introduction 2 Related Works 3 Proposed Method 3.1 Data Selection and Analysis 3.2 Learning Methodology 3.3 Artificial Neural Network 4 Implementation and Evaluation 5 Conclusion and Future Work References Emotion Detection from Text in Social Networks 1 Introduction 2 Emotion Detection 2.1 Rule-Based Approach 2.2 Machine Learning Approach 3 Proposed Approach 4 Experiments and Results 4.1 Identifying Words as Emotional Overtones 4.2 Number of Emotion Categories 4.3 Results of Experiments 5 Conclusions References Finite Libby-Novick Beta Mixture Model: An MML-Based Approach 1 Introduction 2 Model Specification 2.1 Finite Libby-Novick Beta Mixture Model 2.2 Maximum Likelihood and EM Algorithm 2.3 The MML Criterion for a Finite Libby-Novick Beta Mixture 2.4 Fisher Information for Libby-Novick Beta Mixture Model 2.5 Determinant of the Fisher Information 2.6 Prior Distribution 2.7 Full Learning Algorithm 3 Experimental Results 3.1 Malaria Detection 3.2 Breast Tissue Analysis 3.3 Lung Cancer Diagnosis 4 Discussion and Conclusion References Artificial Intelligences on Automated Context-Brain Recognition with Mobile Detection Devices 1 Introduction 2 Related Work 2.1 Brain Detection 2.2 Brain Awareness 2.3 Application 3 Proposed Method 3.1 Primary Concept 3.2 Framework 3.3 Offline Training 3.4 Online Prediction 4 Experiment 4.1 Experimental Setting 4.2 Experimental Result 5 Conclusion References A Novel Meta-heuristic Search Based on Mutual Information for Filter-Based Feature Selection 1 Introduction 2 Related Work 3 Background 3.1 Theoretical Information Based Criteria 3.2 Filter-Based Feature Selection Formulation 4 The Proposed Solution 4.1 The Local Search Strategy 4.2 The Meta-heuristic Search for Feature Selection 5 Experimental Design 5.1 Benchmark Datasets 5.2 Evaluation Plan 5.3 Algorithm Settings 6 Results and Discussion 6.1 Algorithm Performance Analysis 6.2 Algorithm Robustness Analysis 7 Conclusion References Discovering Prevalent Co-location Patterns Without Collecting Co-location Instances 1 Introduction 1.1 Related Work 1.2 Contributions 2 The Basic Concept 3 The New Prevalent Co-location Pattern Mining Framework 3.1 Enumerating Cliques 3.2 Constructing a Co-location Hashmap Structure 3.3 Calculating Participation Indexes and Filtering PCPs 3.4 The Time Complexity Analyses 4 Experiment Evaluations 4.1 Experimental Datasets 4.2 Mining Comparisons 5 Conclusion References Integrating Geospatial Tools for Air Pollution Prediction: A Synthetic City Generator Framework for Efficient Modeling and Visualization 1 Introduction 2 Related Works 2.1 Traditional Computational Models for Air Pollution Prediction 2.2 Technologies Utilized in Computer-Aided Air Pollution Prediction 3 The Synthetic City Generator Framework 4 Discussion 5 Conclusions References Design of an Automated CNN Composition Scheme with Lightweight Convolution for Space-Limited Applications 1 Introduction 2 Related Works 2.1 Neural Architecture Search 2.2 Model Compression 3 Method 3.1 Automatically CNN Composition 3.2 Technical Details 3.3 Efficient Way to Find Network Architecture - Horizontal Calculation 3.4 Rollback Architecture 3.5 A New Kernel Convolution Architecture - Sneaking Feature Compensation Convolution (SFCC) 4 Results 5 Conclusions References Author Index