Database Systems for Advanced Applications: 27th International Conference, DASFAA 2022, Virtual Event, April 11–14, 2022, Proceedings, Part I (Lecture Notes in Computer Science, 13245)

✍ Scribed by Arnab Bhattacharya (editor), Janice Lee Mong Li (editor), Divyakant Agrawal (editor), P. Krishna Reddy (editor), Mukesh Mohania (editor), Anirban Mondal (editor), Vikram Goyal (editor), Rage Uday Kiran (editor)

• Publisher: Springer
• Year: 2022
• Tongue: English
• Leaves: 788
• Category: Library

✦ Synopsis

The three-volume set LNCS 13245, 13246 and 13247 constitutes the proceedings of the 26th International Conference on Database Systems for Advanced Applications, DASFAA 2022, held online, in April 2021.

The total of 72 full papers, along with 76 short papers, are presented in this three-volume set was carefully reviewed and selected from 543 submissions. Additionally, 13 industrial papers, 9 demo papers and 2 PhD consortium papers are included.

The conference was planned to take place in Hyderabad, India, but it was held virtually due to the COVID-19 pandemic.

✦ Table of Contents

General Chairs’ Preface
Program Chairs’ Preface
Organization
Contents – Part I
Contents – Part II
Contents – Part III
Database Queries
Approximate Continuous Top-K Queries over Memory Limitation-Based Streaming Data
1 Introduction
2 Preliminary
2.1 Related Works
2.2 Problem Definition
2.3 The Algorithm S-Merge
3 The Framework -TOPK
3.1 The -MSET
3.2 The Incremental Maintenance Algorithms
3.3 The Optimization Incremental Maintenance Algorithms
4 The Experiment
4.1 Experiment Settings
4.2 The Performance Evaluation
5 Conclusion
References
Cross-Model Conjunctive Queries over Relation and Tree-Structured Data
1 Introduction
2 Preliminary
3 Approach
3.1 Tree and Relational Data Representation
3.2 Challenges
3.3 Cross-Model Join (CMJoin) Algorithm
4 Evaluation
4.1 Evaluation Setup
5 Related Work
6 Conclusion and Future Work
References
Leveraging Search History for Improving Person-Job Fit
1 Introduction
2 Related Work
3 Problem Definition
4 The Proposed Approach
4.1 Text Matching Component
4.2 Intention Modeling Component
4.3 Prediction and Optimization
5 Experiments
5.1 Experimental Setup
5.2 The Overall Comparison
5.3 Evaluation in Different Skill Groups
5.4 Ablation Study
5.5 Performance Tuning
5.6 Qualitative Analysis
6 Conclusion
References
Efficient In-Memory Evaluation of Reachability Graph Pattern Queries on Data Graphs
1 Introduction
2 Preliminaries and Problem Definition
3 Query Reachability Graph
4 A Graph Traversal Filtering Algorithm
5 A Join-Based Query Occurrence Enumeration Algorithm
6 Experimental Evaluation
6.1 Setup
6.2 Performance Results
6.3 Comparison with Graph DB Systems
7 Related Work
8 Conclusion
References
Revisiting Approximate Query Processing and Bootstrap Error Estimation on GPU
1 Introduction
2 Preliminary
2.1 AQP and Bootstrap
2.2 Two Approximate Query Processing Models with GPU
3 AQP and Bootstrap-Based Error Estimation on GPU
3.1 Coprocessor Model
3.2 Main Processor Model
4 Advanced Optimization
4.1 One-Step Calculation
4.2 Count Sampling
4.3 One-Time Hashtable Building
5 Experiments
5.1 Experiment Setup
5.2 Performance Comparison
5.3 Factor Analysis
6 Related Work
7 Conclusion
References
-join: Efficient Join with Versioned Dimension Tables
1 Introduction
2 Join with Multiple Versions of Dimension Tables
3 The -join Operator
4 Evaluation
5 Related Work
6 Conclusion
References
Learning-Based Optimization for Online Approximate Query Processing
1 Introduction
2 Approximate Query Optimization
3 Deep Learning-Based Error Prediction Model
4 Experiment
5 Related Work
6 Conclusion
References
Knowledge Bases
Triple-as-Node Knowledge Graph and Its Embeddings
1 Introduction
2 Related Work
2.1 KGE Datasets
2.2 KGE Techniques
2.3 Event KGs and Representations
3 Problem Formulation
4 Our Model
4.1 E-E Prediction Learning
4.2 F-E Prediction Learning
4.3 Q-E Prediction Learning
5 Dataset
5.1 Dataset Construction
5.2 Conversion Strategies
6 Experiments
6.1 Experimental Setup
6.2 Link Prediction Results
6.3 Analysis
6.4 Case Study
7 Conclusion
References
LeKAN: Extracting Long-tail Relations via Layer-Enhanced Knowledge-Aggregation Networks
1 Introduction
2 Related Works
3 Methodology
3.1 Framework
3.2 Instance Encoder
3.3 Distributed Relational Representation via Transfer Learning
3.4 LKATT
4 Experiments
4.1 Experimental Setting
4.2 Overview of the Evaluation Results
4.3 Ablation Study
4.4 Visualization of Class Embeddings
5 Conclusion
References
TRHyTE: Temporal Knowledge Graph Embedding Based on Temporal-Relational Hyperplanes
1 Introduction
2 Notations
3 Our Model
3.1 Temporal-Relational Hyperplane Projection
3.2 Evolving Modeling
3.3 Dynamic Negative Sampling
3.4 Expand-and-Best-Merge Strategy (Testing phase)
4 Experiments
4.1 Experimental Setup
4.2 Results and Analysis
4.3 Ablation Study and Case Study
5 Related Work
5.1 Static Knowledge Graph Embedding
5.2 Temporal Knowledge Graph Embedding
6 Conclusion
References
ExKGR: Explainable Multi-hop Reasoning for Evolving Knowledge Graph
1 Introduction
2 Related Works
3 Methodology
3.1 Framework of ExKGR
3.2 Emerging Entities Encoder
3.3 Dynamic Reward
3.4 Action Pruning
4 Experiments
4.1 Setup
4.2 Link Prediction Results
4.3 Ablation Study and Analysis
4.4 Qualitative Analysis
5 Conclusion
References
Improving Core Path Reasoning for the Weakly Supervised Knowledge Base Question Answering
1 Introduction
2 Related Work
3 Methodology
3.1 Question Encoding and Path Encoding Module
3.2 Alignment Module with Two-Stage Learning Strategy
4 Experiment
4.1 Experimental Setup
4.2 Results and Analysis
5 Conclusion
References
Counterfactual-Guided and Curiosity-Driven Multi-hop Reasoning over Knowledge Graph
1 Introduction
2 Problem Formulation
3 Methodology
3.1 Path Semantic-Aware Relation Reasoner
3.2 Construct Counterfactuals to Give Soft Rewards
3.3 Intrinsic Curiosity Reward
3.4 Optimization and Training
4 Experiments
5 Conclusion
References
Visualizable or Non-visualizable? Exploring the Visualizability of Concepts in Multi-modal Knowledge Graph
1 Introduction
2 Methodology
2.1 Multi-modal Visualizable Concept Classifier
2.2 Training Under PU Setting
3 Experiment
3.1 Datasets and Settings
3.2 Main Results
3.3 Ablation Study
4 Related Work
5 Conclusion
References
Spatio-Temporal Data
JS-STDGN: A Spatial-Temporal Dynamic Graph Network Using JS-Graph for Traffic Prediction
1 Introduction
2 Related Work
3 Preliminaries
4 Methodology
4.1 Architecture Overview
4.2 JS-Graph Convolution Network
4.3 Dynamic Graph Attention Network
4.4 Spatial Gated Fusion
4.5 Temporal Module
4.6 Other Components
5 Experiments
5.1 Datasets
5.2 Baselines
5.3 Experimental Setup
5.4 Experimental Results
5.5 Study on JS-Graph
5.6 Effectiveness of Each Component
5.7 Case Study
6 Conclusions
References
When Multitask Learning Make a Difference: Spatio-Temporal Joint Prediction for Cellular Trajectories
1 Introduction
2 Related Work
2.1 Trajectory Prediction
2.2 Multitask Learning
3 Problem Definition
4 Our Model
4.1 Overview of IAMT
4.2 Embedding Layer
4.3 Self-attention Layer
4.4 Gating Layer
4.5 Prediction Layer
4.6 Loss Layer
5 Experiments
5.1 Datasets
5.2 Baselines
5.3 Parameter Setup and Metrics
5.4 Comparisons of Performance
6 Conclusion
References
Efficient Retrieval of Top-k Weighted Spatial Triangles
1 Introduction
2 Preliminary
3 Our Solution
4 Experiment
5 Conclusion
References
DIOT: Detecting Implicit Obstacles from Trajectories
1 Introduction
2 Problem Formulation
2.1 Basic Definitions
2.2 Distance Function
2.3 Density Function
2.4 Obstacle Detection
3 DIOT
3.1 The Basic Framework
3.2 Optimizations
4 Experiments
4.1 Quantitative Analysis
4.2 Case Studies
5 Conclusions
References
Exploring Sub-skeleton Trajectories for Interpretable Recognition of Sign Language
1 Introduction
2 Setup and Problem Definition
3 Mining Sub-skeleton Features
4 Experimental Setup and Results
5 Conclusion
References
Significant Engagement Community Search on Temporal Networks
1 Introduction
2 Related Work
3 Significant Engagement Community Search
4 The Top-Down Greedy Peeling Algorithm
5 The Bottom-Up Local Search Algorithm
6 Experimental Evaluation
7 Conclusion
References
Influence Computation for Indoor Spatial Objects
1 Introduction
2 Related Work
2.1 Outdoor Techniques
2.2 Indoor Techniques
3 Preliminaries
3.1 Problem Definition
3.2 Observation
4 IRV Algorithm
4.1 Solution Overview
4.2 Pruning Algorithm
4.3 Verification Algorithm
5 Experimental
5.1 Experimental Settings
5.2 Experiment Results
6 Conclusion
References
A Localization System for GPS-free Navigation Scenarios
1 Introduction
2 System Overview
3 System Deployment
References
Systems
HEM: A Hardware-Aware Event Matching Algorithm for Content-Based Pub/Sub Systems
1 Introduction
2 Related Work
3 Problem Definition
4 Design
4.1 Overview
4.2 Data Structure of HEM
4.3 Matching Procedure of HEM
5 Theoretical Analysis
5.1 Complexity Analysis
5.2 Performance Analysis
6 Experiments
6.1 Setup
6.2 Verification Experiments
6.3 Metric Experiments
6.4 Maintainability
7 Conclusion
References
RotorcRaft: Scalable Follower-Driven Raft on RDMA
1 Introduction
2 Preliminary
2.1 RDMA Network
2.2 Intel Optane DCPMM
2.3 Related Works
3 RotorcRaft Overview
4 Follower-Driven Log Replication
4.1 The Structure of mList
4.2 Mechanism of Follower-Driven Log Replication
4.3 Log Chase
4.4 Log Replication RPC
5 Quorum Follower Read
5.1 Mechanism of Quorum Follower Read
5.2 Follower Read RPC
6 Communication Complexity
7 Evaluation
7.1 Experimental Setup
7.2 Overview Performance
7.3 Log Replication Performance
7.4 Follow Read Performance
7.5 Scalability
8 Conclusion
References
Efficient Matrix Computation for SGD-Based Algorithms on Apache Spark
1 Introduction
2 Motivation
2.1 Motivation for Sampling-Aware Data Loading
2.2 Motivation for Sampling-Aware Data Partition
3 Sampling-Aware Data Loading
3.1 Amount of Redundant IO
3.2 Fine-Grained Data Loading
4 Sampling-Aware Data Partition
4.1 Hash Partition
4.2 Semantic-Based Partition
5 System Implementation
6 Experimental Studies
6.1 Experimental Setting
6.2 Efficiency of Fine-Grained Data Loading
6.3 Efficiency of Semantic-Bases Partition Scheme
7 Related Work
8 Conclusion
References
Parallel Pivoted Subgraph Filtering with Partial Coding Trees on GPU
1 Introduction
2 Partial Coding Tree
3 Partial Adjacency Matrix
4 Experimental Results
4.1 Effect of K
4.2 Comparison with GpSM
5 Related Work
6 Conclusion
References
TxChain: Scaling Sharded Decentralized Ledger via Chained Transaction Sequences
1 Introduction
2 System Overview and Problem Definition
2.1 System Model
2.2 Transaction Model
2.3 Problem Definition
3 Consensus Mechanism in TxChain
3.1 Prerequisites of Transaction Execution
3.2 Transaction Sequence Conversion Algorithm
4 Performance Evaluation
4.1 Experimental Setup
4.2 Throughput Scalability and Transaction Latency of TxChain
5 Conclusion
References
Zebra: An Efficient, RDMA-Enabled Distributed Persistent Memory File System
1 Introduction
2 The Zebra System
2.1 Design
2.2 An Adaptive Replication Transmission Protocol
2.3 Multithreaded RDMA Transmission
3 Evaluation
3.1 Setup
3.2 Sensitivity to I/O Size
3.3 Concurrency
3.4 Scalability
4 Conclusion
References
Data Security
ADAPT: Adversarial Domain Adaptation with Purifier Training for Cross-Domain Credit Risk Forecasting
1 Introduction
2 Related Work
2.1 Domain Adaptation
2.2 Credit Risk Forecasting
2.3 Class-Imbalance
3 Business Setting and Problem Statement
3.1 Business Setting
3.2 Problem Statement
4 The Proposed Model
4.1 The Model
4.2 Multi-source Adversarial Domain Adaptation
4.3 The Training Method
5 Experiments
5.1 Dataset
5.2 Baselines and Compared Methods
5.3 Implementation Details
5.4 Main Results in the CRF Task
5.5 Ablation Test
5.6 Result Visualization
6 Conclusion
References
Poisoning Attacks on Fair Machine Learning
1 Introduction
2 Background
2.1 Fair Machine Learning
2.2 Data Poisoning Attack
3 Data Poisoning Attack on FML
3.1 Problem Formulation
3.2 Convex Relaxation of Fairness Constraint
3.3 Attack Algorithm
4 Experiments
4.1 Evaluation of PFML with Equalized Odds
4.2 Evaluation of PFML with Demographic Parity
4.3 Sensitivity Analysis of Hyperparameters
4.4 Significance Testing
4.5 Summarized Results of Adult Dataset
5 Related Work
6 Conclusions and Future Work
References
Bi-Level Selection via Meta Gradient for Graph-Based Fraud Detection
1 Introduction
2 Methodology
2.1 Instance-level Node Selection
2.2 Neighborhood-level Node Selection
3 Experiments
3.1 Experimental Setup
3.2 Overall Evaluation (RQ1)
3.3 Comparison with Imbalanced Learning Methods (RQ2)
3.4 Ablation Study (RQ3)
4 Related Work
5 Conclusion
References
Contrastive Learning for Insider Threat Detection
1 Introduction
2 Related Work
3 Framework
3.1 Self-supervised Pre-training Component
3.2 Supervised Fine Tuning Component
4 Experiments
4.1 Experimental Setup
4.2 Experimental Results
5 Conclusion
References
Metadata Privacy Preservation for Blockchain-Based Healthcare Systems
1 Introduction
2 Problem Formulation
3 The Proposed Scheme
3.1 Overview
3.2 Construction of Our Scheme
3.3 Security and Privacy Analysis
4 Conclusion and Future Works
References
Blockchain-Based Encrypted Image Storage and Search in Cloud Computing
1 Introduction
2 Related Work
3 Proposed System
4 Theoretical Analysis
5 Performance Evaluations
6 Conclusion
References
Applications of Algorithms
Improving Information Cascade Modeling by Social Topology and Dual Role User Dependency
1 Introduction
2 Related Work
2.1 Diffusion Path Based Methods
2.2 Topological-Based Diffusion Model
3 Methodology
3.1 Problem Definition
3.2 Model Framework
3.3 Embedding Preparation
3.4 Two-Level Attention Networks
3.5 Prediction and Optimization
4 Experiments
4.1 Datasets and Baselines
4.2 Experiment Settings
4.3 Results and Analysis
5 Conclusion
References
Discovering Bursting Patterns over Streaming Graphs
1 Introduction
2 Preliminaries
3 The Baseline Solution
4 A New Approach
4.1 Problem Analysis
4.2 The Progressive Algorithm Framework
4.3 Mapping Subgraphs to Sequences
4.4 Optimization: Edge Sampling
5 Experiments
5.1 Experiments on Different Datasets
5.2 Experiments on Varying Memory
5.3 Experiments on Varying Parameters
6 Related Work
7 Conclusion
References
Mining Negative Sequential Rules from Negative Sequential Patterns
1 Introduction
2 Related Work
2.1 NSP Mining
2.2 Sequential Rule Mining
2.3 NSR Mining
3 Preliminaries
3.1 Positive Sequential Patterns
3.2 Negative Sequential Patterns
4 The nspRule Algorithm
4.1 Review of e-NSP Algorithm
4.2 The Steps of the nspRule Algorithm
4.3 Algorithm Pseudocode
4.4 Analysis of the Time Complexity
5 Experiment with the nspRule Algorithm
5.1 Experiment to Assess the Influence of min_sup
5.2 Experiment to Assess the Influence of min_nor_conf
5.3 Experiment to Assess the Influence of |S|
6 Conclusion
References
CrossIndex: Memory-Friendly and Session-Aware Index for Supporting Crossfilter in Interactive Data Exploration
1 Introduction
2 Preliminaries
2.1 Characterizing Workloads
2.2 Problem Statement
3 Accelerating Crossfilter by CrossIndex
3.1 CrossIndex Construction
3.2 Crossfilter-Style Query Processing
3.3 Optimization for Crossfilter Workloads
4 Experiments
4.1 Setup
4.2 Query Performance
4.3 Offline Cost
4.4 Effect of Construction Order
5 Related Work
6 Discussion
7 Conclusion
References
GHStore: A High Performance Global Hash Based Key-Value Store
1 Introduction
2 Background and Motivation
2.1 Log-Structured Merge Tree
2.2 Motivation
3 GHStore Design
3.1 Global Segmented Hashmap(GHmap)
3.2 GHStore Optimization
3.3 Efficient GHStore Operations
3.4 Crash Consistency
4 Evaluation
4.1 Experimental Setup
4.2 Performance Comparison
4.3 YCSB Workloads
4.4 Performance on SSD
4.5 GHmap Strengths
4.6 Memory Consumption
5 Related Works
6 Conclusion
References
Hierarchical Bitmap Indexing for Range Queries on Multidimensional Arrays
1 Introduction
2 Related Work
3 Preliminaries
3.1 Array Data Model
3.2 Distributed Arrays
3.3 Bitmap Indexing
4 Hierarchical Bitmap Array Index
4.1 Partitioning of Arrays
4.2 Structure of the Array Chunk Index
4.3 Construction of the Hierarchical Bitmap Array Index
4.4 Bin Boundaries Merging in Parent Nodes
4.5 Double Range Encoding of Bitmap Indices in Internal Nodes
4.6 Locality of the Hierarchical Index
5 Querying Dimensions and Attributes
5.1 Attribute Based Matches
5.2 Dimension Based Matches
5.3 Partial and Complete Matches
5.4 Implementation and Fastbit Integration
6 Experimental Evaluation
6.1 Datasets
6.2 Bitmap Indexing Methods
6.3 Range Queries
7 Conclusions and Future Work
References
Membership Algorithm for Single-Occurrence Regular Expressions with Shuffle and Counting
1 Introduction
2 Preliminaries
2.1 SOREs, SOREFCs, MDS and MDC
3 Single-Occurrence Finite Automata with Shuffles and Counters
3.1 Shuffle Markers, Counters and Update Instructions
3.2 Single-Occurrence Finite Automata with Shuffles and Counters
4 Membership Algorithm for SOREFC
5 Experiments
6 Conclusion
References
(p, n)-core: Core Decomposition in Signed Networks
1 Introduction
2 Related Work
3 Problem Statement
4 Algorithms
4.1 Follower-Based Algorithm (FA)
4.2 Disgruntled Follower-Based Algorithm (DFA)
5 Experiments
6 Conclusion
References
TROP: Task Ranking Optimization Problem on Crowdsourcing Service Platform
1 Introduction
2 Problem Statement
3 Offline Task Ranking Optimization
4 Online Task Ranking Optimization
5 Experiments
5.1 The Effectiveness of CTR Vector Prediction
5.2 Performance Comparison
6 Conclusion
References
HATree: A Hotness-Aware Tree Index with In-Node Hotspot Cache for NVM/DRAM-Based Hybrid Memory Architecture
1 Introduction
2 Related Work
3 Hotness-Aware B+-tree
3.1 Index Structure of HATree
3.2 Hotspot Identification
3.3 Operations of HATree
4 Performance Evaluation
4.1 Search Performance
4.2 Updating Performance
5 Conclusions and Future Work
References
A Novel Null-Invariant Temporal Measure to Discover Partial Periodic Patterns in Non-uniform Temporal Databases
1 Introduction
2 Related Work
3 Proposed Model
4 Generalized Partial Periodic Pattern-Growth (G3P-Growth)
5 Experimental Evaluation
6 Conclusions and Future Work
References
Utilizing Expert Knowledge and Contextual Information for Sample-Limited Causal Graph Construction
1 Introduction
2 Preliminaries and Task Definition
3 Methodology
3.1 Phase 1: PU Causal Classifier
3.2 Phase 2: SEM with Subgraphs
4 Experiment
4.1 Experimental Setups
4.2 Experimental Results
5 Conclusion
References
A Two-Phase Approach for Recognizing Tables with Complex Structures
1 Introduction
2 The T2 Framework
2.1 Phase One: Prime Relation Generation
2.2 Phase Two: Graph-Based Alignment Model
3 Experiments
3.1 Experimental Setup
3.2 Evaluation
3.3 Ablation Study
4 Conclusion
References
Towards Unification of Statistical Reasoning, OLAP and Association Rule Mining: Semantics and Pragmatics
1 Introduction
2 Semantic Mapping Between SR and ARM
2.1 Semantic Mapping Between Association Rule Mining and SR (Probability Theory)
2.2 Formal Mapping of ARM Support and Confidence to Probability Theory
3 Semantic Mapping Between SR and OLAP
3.1 Semantic Mapping Between OLAP Averages and SR
4 Conclusion
References
A Dynamic Heterogeneous Graph Perception Network with Time-Based Mini-Batch for Information Diffusion Prediction
1 Introduction
2 Related Work
2.1 Diffusion Path Based Methods
2.2 Social Graph Based Methods
3 Problem Definition
4 Method
4.1 Heterogeneous Graph Construction
4.2 Graph Perception Network (GPN)
4.3 User Dynamic Preferences Based on Mini-Batch
4.4 Dependency-Aware User Embedding
4.5 Fusion Gate
5 Experiments
5.1 Experimental Settings
5.2 Experimental Results
6 Conclusion
References
Graphs
Cascade-Enhanced Graph Convolutional Network for Information Diffusion Prediction
1 Introduction
2 Related Work
2.1 Information Diffusion Prediction
2.2 Graph Neural Networks
3 Problem Statement
4 The Proposed Model
4.1 Cascade-Aware Embedding
4.2 Cascade-Specific Aggregator
4.3 Diffusion Prediction
5 Experimental Setups
5.1 Datasets
5.2 Comparison Methods
5.3 Evaluation Metrics
5.4 Parameter Settings
6 Results and Analysis
6.1 Experimental Results
6.2 Ablation Study
6.3 Parameter Analysis
6.4 Further Study
7 Conclusion
References
Diversify Search Results Through Graph Attentive Document Interaction
1 Introduction
2 Related Work
2.1 Search Result Diversification
2.2 Graph in Search Result Diversification
3 Proposed Model
3.1 Problem Definition
3.2 Architecture
3.3 Diversity Scoring
3.4 Optimization and Ranking
4 Experimental Settings
4.1 Data Collections
4.2 Evaluation Metrics
4.3 Baseline Models
4.4 Implementation Details
5 Experimental Results
5.1 Overall Results
5.2 Discussion and Ablation Study
6 Conclusion
References
On Glocal Explainability of Graph Neural Networks
1 Introduction
2 Related Work
2.1 Local Explanation of GNNs
2.2 Global Explanation of GNNs
3 On the Perspective of Generality
3.1 Counterfactual Qualification
3.2 Candidate Generation
3.3 Mining Strategy
4 On the Perspective of Faithfulness
5 The Proposed Glocal-Explainer
6 Experimental Evaluation
6.1 Datasets and Experimental Setup
6.2 Compared Method
6.3 Candidate Mining Algorithm
6.4 Result and Discussion
7 Conclusion
References
Temporal Network Embedding with Motif Structural Features
1 Introduction
2 Related Work
3 Problem Formulation
4 The Proposed MSTNE Framework
4.1 Neighbor Node Sampling Method Based on the Temporal Motif
4.2 Impact Factor Measure of Temporal Triad
4.3 Attention Mechanisms for Triad with Different Structural Identity and Temporal Relationship
4.4 Loss Function
5 Experimental Results
5.1 Datasets
5.2 Comparison Approaches
5.3 Parameter Settings
5.4 Performance Evaluation
5.5 Desigination of Parameters
6 Conclusion
References
Learning Robust Representation Through Graph Adversarial Contrastive Learning
1 Introduction
2 Methodologies
2.1 Graph Adversarial Attack
2.2 Graph Adversarial Contrastive Learning Framework
3 Theoretical Analysis on Graph Adversarial Contrastive Learning
3.1 Information Bottleneck Principle for Graph Self-supervised Learning
3.2 Generation of Supervised Graph Adversarial Augmentations
3.3 Generation of Unsupervised Graph Adversarial Augmentations
4 Experiments
4.1 Experimental Settings
4.2 Robustness Evaluation Under Netattack
4.3 Robustness Evaluation Under Metattack
4.4 Perturbation Rate Sensitivity for Adversarial Samples
5 Related Work
5.1 Adversarial Attack and Defense on Graph Data
5.2 Self-supervised Graph Representation Learning
6 Conclusion and Discussion
References
What Affects the Performance of Models? Sensitivity Analysis of Knowledge Graph Embedding
1 Introduction
2 Preliminaries: Knowledge Graph Embedding
2.1 General Architecture
2.2 KGE Models
3 A Unified Knowledge Graph Embedding Framework
3.1 Abelian Group and Metric Space
3.2 Group Representation of KGE Models
3.3 Model Transformation and Unification
4 Influencing Factors of Knowledge Graph Models
4.1 Dataset Structural Features
4.2 Embedding Algorithm
4.3 Model Training
5 Sensitivity Analysis of the Influencing Factors in KGE Models
5.1 Experimental Settings
5.2 Sensitivity Analysis of Dataset Structural Features
5.3 Sensitivity Analysis of KGE Model Architecture
5.4 Sensitivity Analysis of Model Training Strategies
6 Conclusion
References
CollaborateCas: Popularity Prediction of Information Cascades Based on Collaborative Graph Attention Networks
1 Introduction
2 Problem Formulation
3 Methodology
3.1 Heterogeneous Bipartite Graph Learning
3.2 Homogeneous Cascade Graph Learning
3.3 Cascade Prediction and Loss Function
4 Evaluation
4.1 Baselines
4.2 Performance Comparison
5 Conclusion
References
Contrastive Disentangled Graph Convolutional Network for Weakly-Supervised Classification
1 Introduction
2 The Proposed Model
2.1 Preliminaries
2.2 Neighborhood Routing Module
2.3 Factor Enhancing Module
2.4 Model Optimization
3 Experiment Evaluation
3.1 Performance Analysis
4 Conclusion
References
CSGNN: Improving Graph Neural Networks with Contrastive Semi-supervised Learning
1 Introduction
2 Related Works
3 Overview
4 Teacher Model with Contrastive Learning
5 Student Model with Reliable Distillation
5.1 Label Reliability Based on Shannon Entropy
5.2 Model Training
6 Experiments
6.1 Experiment Setting
6.2 Semi-supervised Classification
6.3 Ablation Study
7 Conclusion
References
IncreGNN: Incremental Graph Neural Network Learning by Considering Node and Parameter Importance
1 Introduction
2 Related Work
3 Overview of IncreGNN
4 Experience Replay and Regularization Strategy
4.1 Experience Replay Strategy Based on Node Importance
4.2 Regularization Strategy Based on Parameter Importance
5 Experiments
5.1 Experiment Setup
5.2 Experimental Results
6 Conclusion
References
Representation Learning in Heterogeneous Information Networks Based on Hyper Adjacency Matrix
1 Introduction
2 Related Work
3 Preliminaries
4 Method
4.1 Overall Framework
4.2 Node-Level Adjacency Matrix
4.3 Semantic-Level Adjacency Matrix
4.4 Weighted Multi-channel Graph Convolutional Networks
5 Experiment
5.1 Datasets
5.2 Baselines
5.3 Node Classification
5.4 Ablation Study
6 Conclusion
References
Author Index