PRICAI 2022: Trends in Artificial Intelligence: 19th Pacific Rim International Conference on Artificial Intelligence, PRICAI 2022, Shanghai, China, ... Part I (Lecture Notes in Computer Science)

This three-volume set, LNAI 13629, LNAI 13630, and LNAI 13631 constitutes the thoroughly refereed proceedings of the 19th Pacific Rim Conference on Artificial Intelligence, PRICAI 2022, held in Shangai, China, in November 10–13, 2022. The 91 full papers and 39 short papers presented in these volumes were carefully reviewed and selected from 432 submissions. PRICAI covers a wide range of topics in the areas of social and economic importance for countries in the Pacific Rim: artificial intelligence, machine learning, natural language processing, knowledge representation and reasoning, planning and scheduling, computer vision, distributed artificial intelligence, search methodologies, etc. Preface Organization Contents – Part I Contents – Part II Contents – Part III AI Foundations/Decision Theory Fair Allocation with Special Externalities 1 Introduction 2 Preliminaries 3 Reduction from 2-D to 1-D 4 Proportionality in 2-D 5 Approximate MMS in 2-D 5.1 -MMS for Correlated Externality 5.2 -MMS for Inverse Externality 5.3 Re-defining Approximate MMS 6 Conclusion References Robust Weighted Partial Maximum Satisfiability Problem: Challenge to 2P-Complete Problem 1 Introduction 2 Model 3 Complexity of R-MaxSAT Problem 4 R-PMaxSAT Algorithms 4.1 Iterative Best Response (IBR) Algorithm 4.2 Ascending Linear Search in QBF (ALSQ) Algorithm 5 Experimental Evaluation 5.1 Random Instances 5.2 Robust CPP Instances 6 Conclusions References Epistemic Logic via Distance and Similarity 1 Introduction 2 Preliminaries 3 Logics 3.1 Epistemic Logic via Distance (ELD) 3.2 Epistemic Logic via Similarity (ELS) 4 Axiomatic Systems 4.1 Axiomatizations for ELD and ELD- 4.2 Axiomatizations for ELS and ELS- 5 Computational Complexity 5.1 Model Checking 5.2 Satisfiability Checking 6 Conclusion References Abstract Argumentation Goes Quantum: An Encoding to QUBO Problems 1 Introduction 2 Background 2.1 Abstract Argumentation Problems 2.2 QUBO 3 Related Work 4 Encoding 5 Implementation and Tests 5.1 Tests and Comparison 5.2 Execution on the LeapTM Quantum Cloud Service 6 Conclusions and Future Work References Diversification of Parallel Search of Portfolio SAT Solver by Search Similarity Index 1 Introduction 2 Preliminaries 2.1 The SAT Problem and SAT Solvers 2.2 Techniques Employed by SAT Solvers 2.3 Parallelization 2.4 Related Works: Measurement of Distances Between Solvers 3 Methods: Search Similarity Index (SSI) 3.1 Definition of Current Search Direction (CSD) 3.2 Definition of Search Similarity Index (SSI) 3.3 Method to Change Solver's Activity 4 Experiments 5 Concluding Remarks References Dagster: Parallel Structured Search with Case Studies 1 Introduction 2 Related Work 3 System Description 4 Case Study: Costas Arrays 5 Case Study: Pentominoes 6 Case Study: Bounded Model Checking with Abstraction Invariants 7 Conclusions and Future Work References Faster Optimistic Online Mirror Descent for Extensive-Form Games 1 Introduction 2 Related Work 3 Notation and Background 3.1 Treeplex 3.2 Sequence Form 3.3 Dilate Euclidean Distance Generate Function 3.4 Counterfactual Regret Minimization 3.5 Regret Minimization Algorithms 4 Adaptive Methods and Analysis 4.1 Adaptive Regularization Function 4.2 Convergence Analysis 4.3 Dilated Distance Generating Function and Local Minimization 5 Experiment Results and Analysis 5.1 Experimental Setup 5.2 Experimental Results 6 Conclusions References Generalized 3-Valued Belief States in Conformant Planning 1 Introduction 2 Preliminaries 3 Theory 4 Algorithm for Identifying a Base 5 Reduction from Conformant to Classical Planning 5.1 Effects 5.2 Preconditions 5.3 Goals 6 Implementation 6.1 Use of Invariants to Reduce the Base 7 Experiments 8 Related Work 9 Conclusion References Clustering-Based Network Inference with Submodular Maximization 1 Introduction 2 Related Work 3 Methodology 3.1 Problem Formulation 3.2 Proposed Method: CNISM 4 Experimental Evaluation 4.1 Experimental Setup 4.2 Experimental Results on Synthetic Networks 4.3 Experimental Results on Real-World Networks 5 Conclusion References Applications of AI A LiDAR Based Control Solution to Achieve High Precision in Autonomous Parking 1 Introduction 2 Method 2.1 Workflow 2.2 External Factors: Error Feedback System 2.3 Internal Factors: Control Algorithm, Modeling, and Simulation 3 Result 3.1 Apparatus and Testing Scenario 3.2 Experiment and Measurement 3.3 Results 4 Conclusion References Multi-view Heterogeneous Temporal Graph Neural Network for ``Click Farming'' Detection 1 Introduction 2 Related Work 2.1 Graph-Based Anomaly Detection (GBAD) 2.2 Anomaly Detection in Dynamic Graphs 3 Our Framework MHT-GNN 3.1 Overview 3.2 Feature Extraction 3.3 Graph Construction 3.4 Pretraining Phase 3.5 Detection Phase 4 Experiments 4.1 Experiment Setting 5 Conclusion References Deep Forest with Sparse Topological Feature Extraction and Hash Mapping for Brain Network Classification 1 Introduction 2 Related Work 2.1 Random Walk-Based Graph Embedding Methods 2.2 Learning to Hashing 2.3 Extremely Random Forest 3 Deep Forest with Sparse Topological Feature Extraction and Hash Mapping 3.1 Overview of DF-STFEHM 3.2 Sparse Topological Feature Extraction by ERF-WRW 3.3 Sparse Topological Feature Mapping by Kernel Hashing 3.4 CasForest Prediction 4 Experiments 4.1 Datasets and Experimental Settings 4.2 Classification Performance of DF-STFEHM 4.3 Important Brain Regions 5 Conclusions References COVID-19 Forecasting Based on Local Mean Decomposition and Temporal Convolutional Network 1 Introduction 2 Related Work 2.1 Local Mean Decomposition (LMD) 2.2 Temporal Convolutional Network (TCN) 3 Methodology 3.1 Problem Formulation 3.2 Overview of Our Method 3.3 Time Series Decomposition 3.4 Architecture Details 4 Experiments 4.1 Dataset and Preprocessing 4.2 Implementation Details 4.3 Evaluation Metrics 4.4 Experimental Results 5 Conclusion References VMEKNet: Visual Memory and External Knowledge Based Network for Medical Report Generation 1 Introduction 2 Methodology 2.1 The Model Structure 2.2 TF-IDF Embedding Module 2.3 Visual Memory Module 2.4 Parameter Training 3 Experiment Settings 3.1 Dataset 3.2 Baseline and Evaluation Metrics 3.3 Implementation Details 4 Experiment Results and Analyses 4.1 Comparison with Previous Studies 4.2 Qualitative Results and Analyses 4.3 Ablation Studies 5 Conclusion and Future Work References Detecting Video Anomalous Events with an Enhanced Abnormality Score 1 Introduction 2 Related Work 3 Proposal 3.1 Frame Prediction Model 3.2 Appearance Score 3.3 Motion Score 3.4 Dynamic Score and Threshold-Based Integration 4 Experiments 4.1 Datasets and Evaluation Metric 4.2 Implemtation Details 4.3 Comparison with the State-of-the-Art 4.4 Ablation Study 4.5 Qualitative Analysis 4.6 Parameter Dependencies 5 Conclusion References Frequency Domain Based Learning with Transformer for Underwater Image Restoration 1 Introduction 2 Related Work 2.1 Deep Learning for Underwater Image Restoration 2.2 Vision Transformer 2.3 Frequency Domain 3 Methodology 3.1 Network Architecture 3.2 Underwater Residual Transformer Block (URTB) 3.3 Frequency Domain Loss 4 Experiment 4.1 Implementation Details 4.2 Ablation Study 4.3 Comparison with the SOTA 5 Conclusion References MMISeg: A Semi-supervised Segmentation Method Based on Mixup and Mutual Information for Cardiac MRI Segmentation 1 Introduction 2 Related Work 3 Method 3.1 Problem Definition 3.2 Supervised Segmentation 3.3 Unsupervised Segmentation 3.4 Objective Function 4 Experiments and Results 4.1 Dataset and Evaluation Metrics 4.2 Setup and Results 4.3 Analysis of Our Method 4.4 Ablation Experiment 5 Conclusion References Dual-Stream Feature Fusion Network for Detection and ReID in Multi-object Tracking 1 Introduction 2 Method 2.1 Overview of JDE 2.2 Dual-stream Feature Fusion Network 2.3 Multi-scale Cross-Connected Attention Network 2.4 Decoupled Prediction Head and Training Details 2.5 Online Association 3 Experiments 3.1 Implementation Details and Settings 3.2 Ablation Studies 3.3 Comparison with State-of-the-Art MOT Methods 3.4 Qualitative Results 4 Conclusion References A Novel Approach for Pill-Prescription Matching with GNN Assistance and Contrastive Learning 1 Introduction 2 Preliminaries 3 Proposed Method 3.1 Overview 3.2 Pill Detector 3.3 Prescription Recognizer 3.4 Pill-Prescription Alignment 4 Experiments 4.1 Dataset and Experimental Setup 4.2 Experimental Results 4.3 Discussion 5 Conclusion References A Robust Lightweight Deepfake Detection Network Using Transformers 1 Introduction 2 Related Work 3 Approach 3.1 Network Architecture 3.2 Local Feature Extraction 3.3 Robust Transformer 3.4 Spatial Attention Scaling 4 Experiments 4.1 Experiments Setting 4.2 Comparison with the State of the Art 4.3 Ablation Study 4.4 Visualization 5 Conclusion References A General Personality Analysis Model Based on Social Posts and Links 1 Introduction 2 Related Works 3 Data Collection 4 Personality Representation 4.1 Interaction-based Influence Sorting 4.2 Unified Feature Extraction 5 Personality Model Training and Testing 6 Experiment 6.1 Datasets 6.2 Implemention 6.3 Parameters and Metrics 6.4 Baseline Performance 6.5 Impact of High-influence Friend Selection Strategies 6.6 Importing Posts for Users with Limited Posts 7 Conclusion References Deception Detection Towards Multi-turn Question Answering with Context Selector Network 1 Introduction 2 Related Work 2.1 Deception Detection 2.2 Datasets Comparison 3 Model 3.1 Problem Formalization 3.2 Model Overview 3.3 Word Encoder 3.4 Context Selector 3.5 Context Encoder 3.6 Question Answer Pair Classifier 4 Deception QA Dataset Design 4.1 Questionnaires Design 4.2 Answers Collection 4.3 Train/Dev/Test Split 5 Experiments 5.1 Experimental Settings 5.2 Baselines 5.3 Results and Analysis 5.4 Case Study 6 Conclusion References SIA-Unet: A Unet with Sequence Information for Gastrointestinal Tract Segmentation 1 Introduction 2 Methodology 2.1 Sequence Information Processing (SIP) 2.2 Attention Mechanism in U-Net 3 Experiment 3.1 Experimental Settings 3.2 Main Result 3.3 Ablation Study 4 Conclusion References Co-contrastive Self-supervised Learning for Drug-Disease Association Prediction 1 Introduction 2 The Proposed Method 2.1 Multi-source Contrast View Construction 2.2 Context-Aware Neighborhood Aggregation 2.3 Generating Prediction and Model Optimization 2.4 Contrastive Objective 3 Experiments 3.1 Experimental Settings 3.2 Overall Performance 3.3 Model Ablation 4 Conclusion References Obj-SA-GAN: Object-Driven Text-to-Image Synthesis with Self-Attention Based Full Semantic Information Mining 1 Introduction 2 Related Work 3 Object-driven Self-Attention Generative Adversarial Network 3.1 Box Generator 3.2 Shape Generator 4 Experiments 4.1 Setup 4.2 Experimental Results 5 Conclusion and Future Work References Data Mining and Knowledge Discovery APGKT: Exploiting Associative Path on Skills Graph for Knowledge Tracing 1 Introduction 2 Related Work 2.1 Knowledge Tracing 2.2 GNN-based KT Models 2.3 GIKT 3 APGKT: Proposed Model 3.1 Framework 3.2 Graph Construction and Representation 3.3 Student State Evolution and Prediction 4 Experiments 4.1 Setup 4.2 Results 5 Conclusion References Features Fusion Framework for Multimodal Irregular Time-series Events 1 Introduction 2 Related Work 2.1 Multimodal Fusion Problem 2.2 Time-series Forecasting 3 Methodology 3.1 Features Fusion 3.2 Model Architecture 4 Experiments 4.1 Evaluating Metrics 4.2 Comparing Methods 4.3 Experimental Result 5 Conclusion References A Multi-output Integration Residual Network for Predicting Time Series Data with Diverse Scales 1 Introduction 2 Related Work 2.1 Time Series Prediction Methods 2.2 Deep Residual Neural Network 3 Problem Statement 4 Our Proposed Method 4.1 Data Processing 4.2 Residual Module 4.3 Model with Single Residual and Single Predictor 4.4 Integration of Different Prediction Residual Networks 5 Empirical Results 5.1 Dataset 5.2 Experiment Setup and Comparison Baselines 5.3 Prediction Results 5.4 Effect of Residual Structure 5.5 Effect of Different Base Statistical Predictions and Multi-output Structure 6 Conclusion References PLAE: Time-Series Prediction Improvement by Adaptive Decomposition 1 Introduction 2 Related Works 3 Proposed Method 3.1 Predictability Measures 3.2 PLAE 4 Experiments 4.1 Test Data 4.2 PLAE Against Direct Forecasting 4.3 PLAE Against Fixed Decomposition Based Forecasting 5 Conclusion and Future Work References GMEKT: A Novel Graph Attention-Based Memory-Enhanced Knowledge Tracing 1 Introduction 2 Related Work 3 Method 3.1 Overview Architecture 3.2 Embedding Input Module 3.3 Student Knowledge State Update 3.4 Prediction 4 Experiments 4.1 Datasets 4.2 Baseline Model 4.3 Implementation Details 4.4 Experimental Results 4.5 Ablation Study 4.6 Length Analysis and Visualization 5 Conclusion References Dual-VIE: Dual-Level Graph Attention Network for Visual Information Extraction 1 Introduction 2 Related Work 3 Proposed Model 3.1 Token-Level Graph Attention Network 3.2 BD-Level Graph Attention Network 3.3 Reader-Ordered Decoder 4 Experiments 4.1 Datasets 4.2 Implementation Details 4.3 Baseline Method 5 Conclusion References Temporal Edge-Aware Hypergraph Convolutional Network for Dynamic Graph Embedding 1 Introduction 2 Related Work 3 Preliminaries 4 Methodology 4.1 Temporal Hypergraph Construction 4.2 Hyperedge Projection 4.3 Temporal Edge-Aware Hypergraph Convolution 4.4 Loss Function 5 Experiments and Analysis 5.1 Experimental Setup 5.2 Experimental Results 6 Conclusion References Performance Improvement Validation of Decision Tree Algorithms with Non-normalized Information Distance in Experiments 1 Introduction 2 Types of Datasets 3 Disadvantage of ID3 Algorithm 4 Gain Ratio 5 Information Distance-Based Splitting Criteria 5.1 Non-normalized Information Distances (Proposed Methods) 5.2 Advantages of Distances 6 Experiments 6.1 Result Examples for Each Dataset Type 6.2 Advantage of Distances Compared to Gain Ratio 6.3 Comprehensive Experimental Results 7 Conclusion References The Time-Sequence Prediction via Temporal and Contextual Contrastive Representation Learning 1 Introduction 2 Related Work 3 The Proposed PTS-TCC Model 3.1 Cluster Module 3.2 Data Hidden Representation Learning Module 3.3 Temporal Hidden Representation Learning Module 3.4 Contextual Hidden Representation Learning Module 4 Experiment 4.1 Parameters and Preprocessing 4.2 Baseline and Data-Set 4.3 Results and Discussion 5 Conclusions and Future Work References Managing Dataset Shift by Adversarial Validation for Credit Scoring 1 Introduction 2 Dataset Shift 2.1 Definition of Dataset Shift 2.2 Types of Dataset Shift 2.3 Causes of Dataset Shift 3 Methodology 3.1 Adversarial Validation 3.2 Using Adversarial Validation Results to Deal with Dataset Shift 4 Experimental Study 4.1 Data Collection 4.2 Model and Hyperparameters Set-up 4.3 Experiment Set-up 4.4 Results and Discussion 5 Conclusion and Future Work References Linking Check-in Data to Users on Location-aware Social Networks 1 Introduction 2 Related Work 3 Preliminaries 3.1 Variational Autoencoders 3.2 Problem Formulation 4 Proposed Model CULVAE 4.1 Grid Index 4.2 Encoder 4.3 Decoder 4.4 Linking 4.5 Training 5 Experiments 5.1 Datasets 5.2 Compared Methods 5.3 Parameter Settings 5.4 Experimental Results 5.5 Ablation Study 5.6 Parametric Experiment 6 Conclusion and Future Work References Robust Subspace Clustering Based on Latent Low-rank Representation with Weighted Schatten-p Norm Minimization 1 Introduction 2 Related Work 2.1 Low Rank Representation (LRR) 2.2 Latent Low Rank Representation (LLRR) 3 Robust Subspace Clustering Based on Weighted Schatten-p Norm Minimization 3.1 The Proposed Model 3.2 Optimization 3.3 The Complete Clustering Algorithm 3.4 Complexity Analysis 4 Experiments 4.1 Settings 4.2 Experiments on Extended Yale B Dataset 4.3 Experiments on ORL Dataset 4.4 Experiments on COIL-20 Dataset 4.5 Parameter Selection 4.6 Convergence Analysis 5 Conclusion References Evolutionary Computation/Optimisation Speeding up Genetic Programming Based Symbolic Regression Using GPUs 1 Introduction 2 Background and Related Work 2.1 Genetic Programming 2.2 Existing GP Frameworks 3 The Proposed Symbolic Regression Algorithm 3.1 Challenge Faced 3.2 Memory Allocation and Dataset Transfer 3.3 Population Initialization and Selection 3.4 Mutation 3.5 Fitness Evaluation 4 Experiments and Results 4.1 Benchmarks on Synthetic Datasets 4.2 Large-Scale Benchmarks 5 Summary References High-Dimensional Discrete Bayesian Optimization with Intrinsic Dimension 1 Introduction 2 Related Work 3 Preliminaries: Bayesian Optimization 4 Bayesian Optimization with Locality Sensitive Hashing 5 Experiments 5.1 Experimental Setup 5.2 On Synthetic Functions 5.3 On Binary Quadratic Programming Task 6 Conclusion References Multi-objective Evolutionary Instance Selection for Multi-label Classification 1 Introduction 2 Preliminaries and Related Work 2.1 Multi-label k Nearest Neighbor Algorithm 2.2 Instance Selection for ML-kNN 3 Multi-objective Instance Selection for ML-kNN 3.1 Multi-objective Problem Formulation 3.2 MOEIS-ML Algorithm 4 Experimental Studies 4.1 Experimental Setup 4.2 Results and Discussions 5 Conclusion and Future Work References An Investigation of Adaptive Operator Selection in Solving Complex Vehicle Routing Problem 1 Introduction 2 Background 2.1 Stateless Adaptive Operator Selection 2.2 M3CVRP 2.3 Region-Focused Local Search 3 Methodology and Experiment Design 3.1 Experiment 1: AOS in RFLS 3.2 Experiment 2: Neighbourhood Analysis 4 Results and Analysis 4.1 Effectiveness of AOS in RFLS for M3CVRP 4.2 Further Analysis on Neighbours 5 Conclusion References Evolutionary Automated Feature Engineering 1 Introduction 2 Related Work 2.1 Automated Feature Engineering 2.2 Evolutionary Algorithm 3 The Proposed Approach 3.1 Problem Formulation 3.2 Constrained DNA Encoding 3.3 Evolutionary Search Algorithm 4 Experiments 4.1 Experiment Setup 4.2 Effectiveness of EAAFE 4.3 Efficiency of EAAFE 4.4 Effectiveness of the High-Order Transformation 4.5 Generalization Performance of EAAFE 5 Conclusion and Future Work References Author Index