Collaborative Computing: Networking, Applications and Worksharing: 17th EAI International Conference, CollaborateCom 2021, Virtual Event, October 16–18, 2021, Proceedings

Preface
Organization
Contents – Part II
Contents – Part I
Edge Computing
Energy-Efficient Cooperative Offloading for Multi-AP MEC in IoT Networks
	1 Introduction
	2 Related Work
	3 System Model
		3.1 Local Processing Phase
		3.2 Relay MD Processing Phase
		3.3 MEC Server Processing Phase
	4 Problem Formulation and Optimal Solution
		4.1 Energy Efficient Problem and Optimal Solution
		4.2 Supported Maximum Task Length
	5 Numerical Analysis
	6 Conclusions
	References
Multi-truth Discovery with Correlations of Candidates in Crowdsourcing Systems
	1 Introduction
	2 Related Work
		2.1 Truth Discovery
		2.2 Truth Discovery with Different Correlations
		2.3 Multi-truth Discovery
	3 Preliminaries
	4 The Design of MTD-CC
		4.1 Metric for Candidate Correlations
		4.2 Construction of MRF
		4.3 Transformation of MRF
		4.4 Min-cut Based Graph Separation
		4.5 MTD-CC Algorithm
	5 Performance Evaluation
		5.1 Metrics and Baselines
		5.2 Experiment on Real Dataset
		5.3 Experiment on Synthetic Dataset
	6 Conclusion
	References
D2D-Based Multi-relay-Assisted Computation Offloading in Edge Computing Network
	1 Introduction
	2 Motivating Example
	3 System Model and Problem Formulation
		3.1 Local Model
		3.2 Edge Model
		3.3 Incentive Model
		3.4 Computation Offloading Model
	4 D2D-Based Multi-relay-Assisted Computation Offloading Method
		4.1 Relay Selection Algorithm
		4.2 D2D-Based Multi-relay-Assisted Computation Offloading
		4.3 Complexity Analysis
	5 Performance Evaluation
		5.1 Simulation Setting
		5.2 Experimental Results
	6 Related Work and Comparison Analysis
	7 Conclusion
	References
Delay-Sensitive Slicing Resources Scheduling Based on Multi-MEC Collaboration in IoV
	1 Introduction
	2 System Model
		2.1 Slicing Resources Model
		2.2 Mobility and Communication Model
		2.3 Pricing Model
	3 Proposed Solution Strategy
	4 Simulation Results
		4.1 Simulation Settings
		4.2 Parametric Analysis
		4.3 Scheme Comparison
	5 Conclusion
	References
An OO-Based Approach of Computing Offloading and Resource Allocation for Large-Scale Mobile Edge Computing Systems
	1 Introduction
	2 System Model and Problem Formulation
		2.1 Communication Model
		2.2 Computing Model
		2.3 Problem Formulation
	3 An Efficient Offloading Algorithm Based on Ordinal Optimization
		3.1 Computation Resource Allocation Problem
		3.2 Task Placement Problem and An OO-Based Offloading Algorithm
	4 Performance Evaluation
		4.1 Simulation Setup
		4.2 Determine Crude Model and OPC Class
		4.3 Comparison with Baseline Methods
	5 Conclusion
	References
Edge Computing and Collaborative Working
Joint Location-Value Privacy Protection for Spatiotemporal Data Collection via Mobile Crowdsensing
	1 Introduction
	2 Related Work
		2.1 DP-Based Approaches in MCS
		2.2 LDP-Based Approaches in MCS
	3 Motivation
	4 System Model and Preliminaries
		4.1 System Model
		4.2 Preliminaries
	5 Methodology
		5.1 Overview
		5.2 Location Privacy Preserving Mechanism (LPPM)
		5.3 Value Privacy Preserving Mechanism (VPPM)
	6 Theoretical Analysis
	7 Performance Evaluation
		7.1 Simulation Setup
		7.2 Performance Evaluation
	8 Conclusion
	References
Hybrid Semantic Conflict Prevention in Real-Time Collaborative Programming
	1 Introduction
	2 Related Work: Review of Prior Work on Semantic Conflict Prevention with Dependency-Based Automatic Locking (DAL)
	3 Major Constraints of the Prior DAL Scheme
	4 HSCP: Hybrid Semantic Conflict Prevention
	5 Major Technical Issues and Solutions
		5.1 Customizable Locking Scope Determination
		5.2 HSCP Three-Level Awareness Mechanism
		5.3 HSCP Implementation: Architecture and Components
	6 Prototype Implementation and Evaluations
		6.1 Major User Interfaces of HSCP Prototype System
		6.2 Preliminary User Evaluations
		6.3 Experimental Evaluations
	7 Conclusions and Future Work
	References
Supporting Cross-Platform Real-Time Collaborative Programming: Architecture, Techniques, and Prototype System
	1 Introduction
	2 Related Work
	3 Design Objectives and Rationales
		3.1 Design Objective A: Supporting Cross-Platform Real-Time Collaborative Programming
		3.2 Design Objective B: Supporting Unconstrained Multi-level Consistency Maintenance
		3.3 Design Objective C: Supporting Flexible Extensibility and Reusability in Design and Implementation
	4 CP-ROOF: A Novel and Generic Cross-Platform Real-Time Collaborative Programming Framework
		4.1 Workflow and Functional Design
		4.2 Architectural Overview of CP-ROOF
		4.3 CP-ROOF Core: Fundamental Real-Time Collaborative Programming Support
		4.4 CP-ROOF Server: Collaboration Coordinator
		4.5 CP-ROOF Client: Transparent Collaboration Client Adaptor
	5 Major Technical Issues and Solutions
		5.1 Multi-level Operational Transformation
		5.2 Client Design and Implementations
	6 Experimental Evaluations
		6.1 Cross-Platform Collaboration and Evaluations
		6.2 Performance of Major Procedures During Collaboration
	7 Conclusions and Further Work
	References
Collaborative Computing Based on Truthful Online Auction Mechanism in Internet of Things
	1 Introduction
	2 System Model and Problem Formulation
		2.1 Cost Model
		2.2 Utility Model
		2.3 Optimization Problem
	3 Revenue Maximization Online Auction Algorithm Design
		3.1 Pricing Strategy
		3.2 Evaluation of Computation Tasks
		3.3 Online Algorithm
	4 Algorithm Analysis
	5 Experiments Results
	6 Conclusion
	References
A Hashgraph-Based Knowledge Sharing Approach for Mobile Robot Swarm
	1 Introduction
	2 Related Work
		2.1 Robot Swarm
		2.2 Consensus Algorithms in Robot Swarms
		2.3 Hashgraph
	3 Motivated Scenarios
	4 Hashgraph-Based Knowledge Sharing Approach
		4.1 Enhanced Hashgraph in the Mobile Network Environment
		4.2 Hashgraph Approach in the Single-Feature Scenario
		4.3 Hashgraph Approach in the Multi-feature Scenario
	5 Experiments
		5.1 Single-Feature Experiments
		5.2 Multi-feature Experiments
		5.3 Experiments on Consensus Time of Hashgraph Approach
		5.4 Experiments of CPU Utilization
	6 Conclusions and Future Work
	References
Collaborative Working and Deep Learning and Application
CASE: Predict User Behaviors via Collaborative Assistant Sequence Embedding Model
	1 Introduction
		1.1 Top-N Behavior Prediction
		1.2 Limitations of Previous Work
	2 Proposed Model
		2.1 Constructing Graphs
		2.2 Learn Behavior Embedding
		2.3 Predict the Future Behavior
	3 Experiments
		3.1 Datasets
		3.2 Evaluation Metric
		3.3 Experiment Design
		3.4 Result and Analysis
	4 Conclusion
	References
A Collaborative Optimization-Guided Entity Extraction Scheme
	1 Introduction
	2 Related Work
		2.1 The Rule and Vocabulary-Based Entity Extraction
		2.2 The Traditional Machine Learning-Based Entity Extraction
		2.3 The Deep Learning-Based Entity Extraction
	3 The Proposed Scheme
		3.1 The Use of BERT Model
		3.2 The Design of CRF Layer
		3.3 The PSO-Based Collaborative Optimization
	4 Experiments
		4.1 Experimental Dataset
		4.2 Metric
		4.3 Experimental Comparison
	5 Conclusion
	References
A Safe Topological Waypoints Searching-Based Conservative Adaptive Motion Planner in Unknown Cluttered Environment
	1 Introduction
	2 Related Works
		2.1 Standard Pathfinding Algorithm
		2.2 Trajectory Replanning
	3 Geometrically Topological Waypoints Searching
		3.1 Topological Points Searching
		3.2 Optimal State Transition Waypoint
	4 Conservative Trajectory Replanning
		4.1 Adaptive Trajectory Replanning
		4.2 B-Spline Trajectory Representation
		4.3 Problem Formulation
	5 Experiments
		5.1 Implementation Details
		5.2 Topological Waypoints Searching
		5.3 Conservative Trajectory Replanning
		5.4 Comparisons of Planning Efficiency
	6 Conclusions
	References
Multi-D3QN: A Multi-strategy Deep Reinforcement Learning for Service Composition in Cloud Manufacturing
	Abstract
	1 Introduction
	2 Related Work
		2.1 Meta-heuristics-based Service Composition
		2.2 RL/DRL-Based Service Composition
	3 Service Composition Problem Description and MDP-Based CMfg Service Composition
		3.1 Problem Description
		3.2 MDP-Based CMfg Service Composition
	4 Proposed Algorithm Framework
		4.1 DQN Algorithm
		4.2 The Dueling Architecture
		4.3 The Double Estimator
		4.4 The Prioritized Replay Mechanism
		4.5 The Strategy for Adaptability
	5 Experiments
		5.1 Experiment Setting
		5.2 Result Analysis
	6 Conclusions
	Acknowledgments
	References
Transfer Knowledge Between Cities by Incremental Few-Shot Learning
	Abstract
	1 Introduction
	2 Related Work
		2.1 Spatial-Temporal Prediction
		2.2 Transfer Learning
	3 Preliminaries
	4 Methodology
		4.1 The Spatial-Temporal Network Architecture (ST-Net)
		4.2 Incremental Few-Shot Learning
		4.3 Memory Regularizer Combine Base and Novel Knowledge
		4.4 Parameter Optimization (Knowledge Transfer)
	5 Experiment
		5.1 Datasets
		5.2 Data Preprocessing
		5.3 Baselines
		5.4 Evaluation Metric
		5.5 Experimental Settings
		5.6 Experimental Result
		5.7 Parameter Sensitivity
	6 Conclusions
	Acknowledgement
	References
Deep Learning and Application
Multi-view Representation Learning with Deep Features for Offline Signature Verification
	1 Introduction
	2 Related Work
		2.1 Feature Extraction in Offline Signature Verification Systems
		2.2 Multi-view Representation Learning
	3 Multi-view Representation Learning for Offline Signature Verification Systems
		3.1 Generating the Second View from Deep Features
		3.2 CCA-based Multi-view Representation Learning Approaches
		3.3 Training the Writer-Dependent Classifiers
	4 Experiment
		4.1 Dataset
		4.2 Experimental Settings
		4.3 Evaluation Measurements
		4.4 Experiments on the GPDS Dataset
		4.5 Experiments on the CEDAR and Brazilian PUC-PR Datasets
	5 Conclusion
	References
Backdoor Attack of Graph Neural Networks Based on Subgraph Trigger
	Abstract
	1 Introduction
	2 Related Works
	3 Preliminaries
		3.1 Concepts
		3.2 Attack Model
	4 Backdoor Attack on GNN
		4.1 Attack Overview
		4.2 Generation of Trigger
		4.3 Selection of Attack Nodes
		4.4 Backdoor Insertion
	5 Experiment and Evaluation
		5.1 Datasets
		5.2 Evaluation Metrics
		5.3 Experimental Setup
		5.4 Result and Analysis
	6 Conclusions and Future Work
	Acknowledgement
	References
A UniverApproCNN with Universal Approximation and Explicit Training Strategy
	1 Introduction
	2 Preliminaries
	3 A CNN Structure with Universal Approximation: UniverApproCNN
		3.1 Model Design
		3.2 UniverApproCNN for Multiple Outputs
		3.3 Normalized CNN and UniverApproCNN
	4 UniverApproCNN for Two-Dimensional Input
		4.1 Proof of the Approximation of the CNN Suitable for Two-dimensional Input
		4.2 Model Design
	5 Performance Experiment of UniverApproCNN for Inertial Guidance
	6 Approximation Coefficients of UniverApproCNN
	7 Conclusion
	A  Appendix
		A.1 Proof of the Theorem 3
		A.2 Model Training Results in the Trajectory Prediction Experiments
		A.3 Back Propagation Process of UniverApproCNN
	References
MS-BERT: A Multi-layer Self-distillation Approach for BERT Compression Based on Earth Mover's Distance
	1 Introduction
	2 Preliminaries
		2.1 Self-distillation
		2.2 Sample-wise Adaptive Inference
	3 Methodology
		3.1 Overview of MS-BERT
		3.2 Self-distillation with Earth Mover's Distance
		3.3 Student Classifiers Splicing Strategy
		3.4 Top-K Uncertainty
	4 Experiments
		4.1 Datasets
		4.2 Evaluation Metrics
		4.3 Experimental Result
		4.4 Ablation Study
	5 Related Work
	6 Conclusion
	References
Smart Contract Vulnerability Detection Based on Dual Attention Graph Convolutional Network
	1 Introduction
	2 Related Works
	3 Problem Description
	4 Smart Contract Vulnerability Detection Method
		4.1 Attributes Generation
		4.2 Construction of DA-GCN Model
	5 Experiment
		5.1 Experimental Settings
		5.2 Experimental Results
	6 Conclusion
	References
Crowdturfing Detection in Online Review System: A Graph-Based Modeling
	Abstract
	1 Introduction
	2 Related Work
	3 Crowdturfing Detection Model
		3.1 Modeling and Initialization
		3.2 CrowdDet Framework
		3.3 Improved Cost-Sensitive Loss Function
	4 Experiment
		4.1 Dataset and Evaluation Metrics
		4.2 Baseline Methods
		4.3 Data Preparation and Experiment Settings
		4.4 Performance Evaluation
	5 Conclusion
	Acknowledgement
	References
Attention-Aware Actor for Cooperative Multi-agent Reinforcement Learning
	1 Introduction
	2 Background
		2.1 Attention Mechanism
		2.2 Attention-Based Algorithms for MARL
		2.3 Graph Network
	3 Our Approach
		3.1 Multi-agent Mutual Interplay Graph Structure
		3.2 Attention-Aware Actor Architecture
	4 Experimental Evaluation
		4.1 Settings
		4.2 Validation
		4.3 Attention Analysis
	5 Conclusion
	References
Geographic and Temporal Deep Learning Method for Traffic Flow Prediction in Highway Network
	1 Introduction
	2 Related Work
	3 Feature Pre-processing and Prediction Method
		3.1 Feature Engineering
		3.2 Graph Construction
		3.3 Traffic Flow Prediction
	4 Evaluation
		4.1 Settings
		4.2 Experiments
	5 Conclusion
	References
How are You Affected? A Structural Graph Neural Network Model Predicting Individual Social Influence Status
	1 Introduction
	2 Related Work
		2.1 Social Influence
		2.2 Graph Neural Networks
	3 Preliminaries
	4 Our Model: SGN
		4.1 Friendship Interacting Module
		4.2 Influence Propagating Module
		4.3 Global Attention and Output
	5 Experiment
		5.1 Experiment Setings
		5.2 Experiment Result of Prediction Performances
		5.3 Attention Analysis
	6 Conclusion
	References
Multi-order Proximity Graph Structure Embedding
	1 Introduction
	2 Related Work
	3 Proposed Method
		3.1 Notation
		3.2 Structure Preserving Model
		3.3 Random Walk Sampling
		3.4 Negative Sampling
		3.5 Algorithm
	4 Evaluation
		4.1 Experimental Setup
		4.2 Experimental Results
		4.3 Ablation Study of Negative Sampling
		4.4 Study of Hyper Parameters
	5 Conclusion
	References
Deep Learning and Application and UVA
PATR: A Novel Poisoning Attack Based on Triangle Relations Against Deep Learning-Based Recommender Systems
	1 Introduction
	2 Related Work
	3 Proposed Model
		3.1 Problem Formulation
		3.2 Pre-training Module
		3.3 Reconstruction Module
	4 Experiments and Analysis
		4.1 Experimental Setup
		4.2 Experimental Results and Analysis
	5 Conclusion
	References
Low-Cost LiDAR-Based Vehicle Detection for Self-driving Container Trucks at Seaport
	1 Introduction
	2 Related Work
		2.1 LiDAR 3D Point Cloud Detection
		2.2 LiDAR Projection-Based Detection
	3 Dual-LiDAR Perceptive System
		3.1 Data Collection and BEV Map Projection
		3.2 Lightweight CNN Detector
	4 Experimental Study
		4.1 Data Augmentation
		4.2 Comparison of Models
		4.3 Tracking Test
	5 Conclusion
	References
Author Index