PRICAI 2023: Trends in Artificial Intelligence: 20th Pacific Rim International Conference on Artificial Intelligence, PRICAI 2023, Jakarta, Indonesia, ... I (Lecture Notes in Artificial Intelligence)

Preface
Organization
Contents – Part I
Contents – Part II
Contents – Part III
Agents/Decision Theory
DAGE: Dropout with Action Gradient Estimator for Continuous Control
	1 Introduction
	2 Related Work
	3 Background
	4 Estimate Action Gradient Accurately
		4.1 You Need to Minimize Action Gradient Error
		4.2 Dropout Operator for Consistent Bellman Update
	5 DAGE Framework
	6 Experiment
		6.1 Overall Performance
		6.2 Does DAGE Estimate Action Gradient More Accurately?
		6.3 Ablation Study
		6.4 Parameter Sensitivity
	7 Conclusion
	References
Conditional Variational Inference for Multi-modal Trajectory Prediction with Latent Diffusion Prior
	1 Introduction
	2 Methodology
		2.1 Preliminaries and Definitions
		2.2 Conditional Variational Inference with Latent Diffusion Prior
		2.3 Sampling with Classifier-Free Guidance
		2.4 Training Objective and Model Design
	3 Experiments
		3.1 Benchmark Results
		3.2 Ablation Study
	4 Conclusion
	References
Egalitarian Price of Fairness for Indivisible Goods
	1 Introduction
	2 Results
	References
Intelligent Network Intrusion Detection and Situational Awareness for Cyber-Physical Systems in Smart Cities
	1 Introduction
	2 Related Work
	3 Construction of Network Situation Awareness System
		3.1 Overall System Architecture
		3.2 Intelligent Sense Module Construction
		3.3 Flow Packet Capture and Parsing
		3.4 System Visualization
	4 Experiment Results and Analysis
	5 Summarize
	References
Data Mining and Knowledge Discovery
A Dynamic Linear Bias Incorporation Scheme for Nonnegative Latent Factor Analysis
	1 Introduction
	2 Methodology
		2.1 Problem Formulation
		2.2 The Proposed Dynamic Linear Biases
		2.3 The Proposed DNLFA Model
		2.4 Algorithm Design and Analysis
	3 Experiments
		3.1 Experimental Setup
		3.2 Effects of Hyper-Parameter on DNLFA’s Performance
		3.3 Comparison Results
	4 Conclusion
	References
An Anomaly Detection Framework for System Logs Based on Ensemble Learning
	1 Introduction
	2 Related Work
	3 Approach
		3.1 Overall Framework
		3.2 Data Preprocessing
		3.3 Models Training
		3.4 Anomaly Detection with Ensemble
	4 Experiment and Analysis
		4.1 Experimental Setup
		4.2 Results and Analysis
	5 Conclusion
	References
Be Informed of the Known to Catch the Unknown
	1 Introduction
	2 Extant Works
	3 Approach
	4 Empirical Setup
	5 Results and Analysis
	6 Conclusion
	References
Network Structure Embedding Method Based on Role Domain Feature
	1 Introduction
	2 Related Work
	3 Methodology
		3.1 Notions
		3.2 Overall Framework
		3.3 Feature Extraction
		3.4 Representation Learning
	4 Experiments
		4.1 Datasets and Baselines
		4.2 Experiment Settings
		4.3 Experiments on Role Classification
		4.4 Visualization
		4.5 Case Study: Role Discovery
	5 Conclusion and Further Discussion
	References
Machine Learning-Driven Reactor Pressure Vessel Embrittlement Prediction Model
	1 Introduction
	2 Methodology
		2.1 Proposed VPMLP Model
		2.2 Framework of VAE Model
		2.3 Physical Formula Guided Multilayer Perceptron
	3 Experiments
		3.1 Dataset, Evaluation Metrics and Baselines
		3.2 Experimental Results
	4 Conclusion
	References
Mitigating Misinformation Spreading in Social Networks via Edge Blocking
	1 Introduction
		1.1 Preliminaries
		1.2 Prior Work
	2 Proposed Algorithm
	3 Evaluation
		3.1 Comparison of Algorithms
	References
Multi-modal Component Representation for Multi-source Domain Adaptation Method
	1 Introduction
	2 Proposed Method
		2.1 Semantic Representation of Class Components
		2.2 Multi-modal Invariant Representation Learning
	3 Experiments
		3.1 Datasets and Baselines
		3.2 The Performance
	4 Conclusion
	References
(Deep) Reinforcement Learning
Abbreviated Weighted Graph in Multi-Agent Reinforcement Learning
	1 Introduction
	2 Related Works
	3 Background
		3.1 Decentralized POMDP
		3.2 Value Decomposition and CTDE Paradigm
		3.3 GCN and Its Attentional Applications in MARL
	4 Methods
		4.1 Attribution Module
		4.2 Abbreviated Weighted Graph Module
	5 Experiments
		5.1 Results
		5.2 Ablations
	6 Conclusion
	References
Diverse Policies Converge in Reward-Free Markov Decision Processes
	1 Introduction
	2 Related Work
	3 Preliminaries
	4 Methodology
		4.1 A Unified Framework for Diversity Algorithms
		4.2 Convergence Analysis
		4.3 A Contextual Bandit Formulation
		4.4 Regret Bound
	5 Experiments
		5.1 A Geometric Perspective on Policy Evolution
		5.2 Policy Selection Ablation
	6 Conclusion
	References
PruVer: Verification Assisted Pruning for Deep Reinforcement Learning
	1 Introduction
	2 Related Work
		2.1 Pruning in Deep Reinforcement Learning
		2.2 DRL Network Verification
	3 Methodology
		3.1 Refinement of the Pruned Network \'
	4 Case Studies
	5 Conclusions
	References
AdaptLight: Toward Cross-Space-Time Collaboration for Adaptive Traffic Signal Control
	1 Introduction
	2 Method
		2.1 Spatial-Temporal Graph Transformer Network
		2.2 AR-MADRL Framework for Heterogeneous Decision
	3 Experiments
	4 Conclusion
	References
DeepLRA: An Efficient Long Running Application Scheduling Framework with Deep Reinforcement Learning in the Cloud
	1 Introduction
	2 Method
		2.1 Deep Reinforcement Learning Method
		2.2 Model Training
	3 Evaluation
		3.1 Experimental Setup
		3.2 Scheduling Results
	4 Conclusion
	References
Guiding Task Learning by Hierarchical RL with an Experience Replay Mechanism Through Reward Machines
	1 Introduction
	2 HOERM Through Reward Machines
		2.1 Reward Machine
		2.2 Proposed HOERM
	3 Experiments
		3.1 Experimental Setup
		3.2 Experiment 1: Results in Minecraft
		3.3 Experiment 2: Results in Water World
	4 Discussion
	References
Generative AI
A Semantic Similarity Distance-Aware Contrastive Learning for Abstractive Summarization
	1 Introduction
	2 Proposed Method
		2.1 Problem Definition
		2.2 Semantic Similarity Contrastive Learning
		2.3 Semantic Similarity Distance-Aware Contrastive Learning
	3 Experimental Results and Analysis
		3.1 Datasets and Evaluation Metrics
		3.2 Implementation Details
		3.3 Main Results
		3.4 Analysis of Sentence Salience
		3.5 Selection of Positive and Negative Samples
		3.6 Case Study
	4 Conclusions
	References
CCDWT-GAN: Generative Adversarial Networks Based on Color Channel Using Discrete Wavelet Transform for Document Image Binarization
	1 Intorduction
	2 Related Work
	3 Proposed Method
		3.1 Image Preprocessing
		3.2 Image Enhancement
		3.3 Image Binarization
		3.4 Loss Function
	4 Experiments
		4.1 Datasets
		4.2 Evaluation Metric
		4.3 Experiment Setup
		4.4 Ablation Study
		4.5 Experimental Results
	5 Conclusion
	References
Coarse-to-Fine Response Generation for Document Grounded Conversations
	1 Introduction
	2 Related Work
	3 Methodology
		3.1 Task Definition
		3.2 Coarse-Grained Feature Extraction Module
		3.3 Fine-Grained Feature Generation Module
	4 Experiment
		4.1 Dataset
		4.2 Evaluation Metrics
		4.3 Baselines
		4.4 Implementation Details
		4.5 Experimental Results
		4.6 Ablation Study
	5 Conclusion
	References
Context-Dependent Text-to-SQL Generation with Intermediate Representation
	1 Introduction
	2 Related Work
	3 Methodology
		3.1 Problem Setup
		3.2 Intermediate Representation
		3.3 Model
	4 Experiments
		4.1 DataSet and Metrics
		4.2 Baseline Models
		4.3 Implementation Details
		4.4 Experiment Results
		4.5 Ablation Study
	5 Conclusion
	References
CSS: Contrastive Span Selector for Multi-span Question Answering
	1 Introduction
	2 Related Work
		2.1 Tagger Model in MSQA
		2.2 Span-Based Method in Other Extractive Task
	3 Methodology
		3.1 Task Definition
		3.2 Model Overview
		3.3 Contrastive Learning for MSQA
	4 Experiments
		4.1 Experiments Settings
		4.2 Main Results
		4.3 Ablation Study
	5 Conclusion and Future Work
	References
Generative Model of Suitable Meme Sentences for Images Using AutoEncoder
	1 Introduction
	2 GUMI-AE: Generative Model of Suitable Meme Sentences for Images Using AutoEncoder
		2.1 Image Caption Generative Model
		2.2 AutoEncoder
		2.3 GUMI-AE
	3 Experimental Results
		3.1 Dataset for Training
		3.2 Training of AutoEncoder
		3.3 Training of Meme Generator
		3.4 Evaluation of Humorous
	4 Conclusion
	References
MTMG: A Framework for Generating Adversarial Examples Targeting Multiple Learning-Based Malware Detection Systems
	1 Introduction
	2 Related Work
	3 Design and Implementation
		3.1 MTMG Overview
		3.2 Q1\'s Solution
		3.3 Q2\'s Solution
		3.4 Training Algorithm
	4 Evaluation
		4.1 Attack on Single LB-MDS
		4.2 Attack on Multiple LB-MDS
		4.3 Algorithm Efficiency Evaluation
	5 Conclusion
	References
Semantic Segmentation of Remote Sensing Architectural Images Based on GAN and UNet3+ Model
	1 Introduction
	2 Related Work
		2.1 Deep Neural Network
		2.2 Generative Adversarial Network
		2.3 Attention Mechanism
	3 Research Method
		3.1 Network Structure
		3.2 Loss Function
	4 Research
		4.1 Dataset
		4.2 Evaluation Metrics
		4.3 Experimental Setups
		4.4 Experimental Results and Analysis
	5 Conclusion and Future Work
	References
Sparse Reconstruction Method for Flow Fields Based on Mode Decomposition Autoencoder
	1 Introduction
		1.1 Contributions
	2 Problem Definition of Sparse Reconstruction
	3 Framework of Model
		3.1 Training Method of Mode Decomposition Autoencoder
		3.2 Refactoring Method
		3.3 Mode Decomposition Autoencoder Models
	4 Results and Discussion
		4.1 Example 1: NOAA Sea Surface Temperature
		4.2 Example 2: Dimensional Multi-cylinder Wake
		4.3 Discussion
	5 Conclusions
	References
StyleDisentangle: Disentangled Image Editing Based on StyleGAN2
	1 Introduction
	2 Related Works
		2.1 Latent Space Manipulation
		2.2 Text-Based Image Manipulation
	3 Methodology
		3.1 Overview
		3.2 Attribute Coordinates
		3.3 Objective Function
	4 Experiments
		4.1 Experimental Setup
		4.2 StyleDisentangle Manipulation Results
		4.3 Comparison with Text-Guided Methods
		4.4 Ablation Studies
	5 Conclusion
	References
A Property Constrained Video Summarization Framework via Regret Minimization
	1 Introduction
	2 The Property Constrained Video Summarization Framework via Regret Minimization
		2.1 Constructing the Candidate Frame Set
		2.2 Transforming to a Multi-dimensional Point Set
		2.3 Generating Keyframes via the Regret Minimization Query
		2.4 Adding Keyframes by Storyness
	3 Experiments
	4 Conclusion and Future Work
	References
Enhancing Keyphrase Generation by BART Finetuning with Splitting and Shuffling
	1 Introduction
	2 Related Work
	3 Keyphrase-Focused BART
	4 Experiments
		4.1 Experimental Settings
		4.2 Results and Analysis
	5 Conclusion and Future Work
	References
Graph Learning
A Dynamic-aware Heterogeneous Graph Neural Network for Next POI Recommendation
	1 Introduction
	2 Related Work
	3 Problem Formulation
	4 Method
		4.1 Dynamic-Aware Heterogeneous Graphs Construction
		4.2 Fine-Grained Temporal Enhanced Graph Neural Network
		4.3 Dynamic Information Aggregation Module
	5 Experiments
		5.1 Datasets and Preprocessing
		5.2 Evaluation Metrics
		5.3 Baseline Models
		5.4 Parameter Settings
		5.5 Performance Comparisons
		5.6 Ablation Study
		5.7 Hyperparameter Analysis
	6 Conclusion
	References
Cross-scale Dynamic Relation Network for Object Detection
	1 Introduction
	2 Method
		2.1 Multi-scale Features Extraction
		2.2 Cross-scale Semantic-Aware Module
		2.3 Dynamic Relation Graph Reasoning
		2.4 Semantic Attention Fusion Module
	3 Experimental Results and Analysis
		3.1 Dataset and Evaluation Metrics
		3.2 Implementation Details
		3.3 Main Results
		3.4 Ablation Studies
		3.5 Qualitative Analysis
	4 Conclusion
	References
Distribution-Adaptive Graph Attention Networks for Flood Forecasting
	1 Introduction
	2 Related Work
		2.1 Data-Driven Flood Prediction Models
		2.2 Application of Transfer Learning in Time Series
		2.3 Graph Attention Networks
	3 Methodology
		3.1 Hydrological Spatial Homogeneous Graph Generation Module
		3.2 Hydrological Distribution Characterization Module
		3.3 Spatio-Temporal Graph Attention Networks Module
		3.4 Distribution Adaptive Module
	4 Experiments
		4.1 DataSet and Measurements
		4.2 Implementation Details
		4.3 Performance Comparison
		4.4 Performance Analysis
	5 Conclusion
	References
DSAM-GN: Graph Network Based on Dynamic Similarity Adjacency Matrices for Vehicle Re-identification
	1 Introduction
	2 Related Work
		2.1 CNNs and Graph Networks
		2.2 Node and Edge Construction in GNs for Vehicle Re-ID
	3 Proposed Method
		3.1 Overview
		3.2 DSAM-GN
		3.3 Loss Function
	4 Experiments
		4.1 Implementation Details
		4.2 Experimental Results and Analysis
		4.3 Ablation Study
	5 Conclusion
	References
Dynamic Spatial-Temporal Dual Graph Neural Networks for Urban Traffic Prediction
	1 Introduction
	2 Preliminaries and Problem Definition
		2.1 Notations and Symbols
		2.2 Problem Definition and Description
	3 Methodology
		3.1 Dynamic Spatiotemporal Graph Construction
		3.2 Spatial-Temporal Attention Module
		3.3 Spatial-Temporal Convolution Module
	4 Experiments
		4.1 Datasets Used in the Experiment
		4.2 Baseline Used in the Experiment
		4.3 Setup of Experiments
		4.4 Experimental Analysis
		4.5 Ablation Experiments
	5 Conclusion
	References
MuHca: Mixup Heterogeneous Graphs for Contrastive Learning with Data Augmentation
	1 Introduction
	2 Related Work
	3 Methodology
		3.1 Graph Data Augmentation
		3.2 Contrastive Loss
	4 Experiments and Results
		4.1 Datasets
		4.2 Baselines
		4.3 Experimental Settings
		4.4 Node Classification Results
		4.5 Performance Comparison
		4.6 Parameter Sensitivity
	5 Conclusion
	References
Parameter-Lite Adapter for Dynamic Entity Alignment
	1 Introduction
	2 Related Work
		2.1 Static Entity Alignment
		2.2 Dynamic Entity Alignment
	3 Methodology
		3.1 Overview
		3.2 Mapping-Based Feature Fusion
		3.3 Parameter-Lite Adapter Tuning
	4 Experiments
		4.1 Experimental Setup
		4.2 Main Results
		4.3 Ablation Study
	5 Conclusion
	References
Zoom-Based AutoEncoder for Origin-Destination Demand Prediction
	1 Introduction
	2 Related Work
		2.1 Origin-Destination Demand Prediction
		2.2 Autoencoder
	3 Problem Formulation
		3.1 Definitions
		3.2 Problem Definition
	4 Methodology
		4.1 Motivation and Overview
		4.2 Zoom Based Encoder
		4.3 Training Strategy
	5 Experiment
		5.1 Datasets
		5.2 Baselines and Metrics
		5.3 Experimental Settings
		5.4 Comparison Results
		5.5 Ablation Study
	6 Conclusion
	References
Healthcare and Wellbeing
A Stagewise Deep Learning Framework for Tooth Instance Segmentation in CBCT Images
	1 Introduction
	2 Methodology
		2.1 Coarse Segmentation and ROI Extraction Stage
		2.2 Fine Segmentation Stage
		2.3 Potential Energy Loss
	3 Experiment
		3.1 Data Preprocessing
		3.2 Training and Results
	4 Conclusion and Discussion
	References
Hierarchical Pooling Graph Convolutional Neural Network for Alzheimer\'s Disease Diagnosis
	1 Introduction
	2 Related Work
		2.1 Graph Convolutional Neural Networks
		2.2 Graph Classification Tasks
		2.3 Graph Pooling
	3 Method
		3.1 Graph Construction
		3.2 Graph Hierarchical Pooling
	4 Experiments
		4.1 Data Acquisition and Preprocessing
		4.2 Experimental Setup
		4.3 Hyperparameter Settings
		4.4 Analysis of Experimental Results
	5 Conclusion
	References
Learning Cross-Modal Factors from Multimodal Physiological Signals for Emotion Recognition
	1 Introduction
	2 Related Work
	3 Proposed Approach
		3.1 Updating the Hidden States
		3.2 Model Training
	4 Experiment
		4.1 Dataset
		4.2 Step1 - Extraction of Emotion Features
		4.3 Step 2 and 3 - Feature Selection and Linear Regression
		4.4 Ablation Study
	5 Results and Discussion
	6 Conclusions, Limitations and Future Direction
	References
VaeSSC: Enhanced GRN Inference with Structural Similarity Constrained Beta-VAE
	1 Introduction
	2 Method
		2.1 Generalized Structural Equation Modeling
		2.2 Loss Function Design
		2.3 VaeSSC Framework
		2.4 Ensemble Strategy
	3 Experiments and Results
		3.1 Data Preprocessing
		3.2 Experimental Result
	4 Conclusion
	References
Knowledge Representation and Reasoning
Parallel Construction of Knowledge Graphs from Relational Databases
	1 Introduction
	2 Preliminaries
	3 The Fingr Engine
		3.1 R2RML Mapping Resolution
		3.2 Triple Generation
		3.3 Triple Storage
		3.4 Overview
	4 Experiments
		4.1 Evaluation over the GTFS Benchmark
		4.2 Evaluation over the Berlin SPARQL Benchmark
	5 Conclusion and Discussion
	References
Knowledge Graph Augmentation with Entity Identification for Improving Knowledge Graph Completion Performance
	1 Introduction
	2 Related Work
	3 Entity Identification Based on Graph Information
		3.1 Augmentation with Entity Identification
		3.2 Feature Vectors of Entities with BERT Considering Graph Information
	4 Experiments and Evaluations
		4.1 Settings
		4.2 Results and Discussion
	5 Conclusion
	References
Relational Acceptability Semantics of Abstract Argumentation
	1 Introduction
	2 Technical Preliminaries
	3 Argumentation Tuple Relational Calculus and Relational Acceptability Semantics
	4 Conclusions
	References
Author Index