Machine Learning for Cyber Security: 4th International Conference, ML4CS 2022, Guangzhou, China, December 2–4, 2022, Proceedings, Part II

Preface
Organization
Contents – Part II
AMAD: Improving Adversarial Robustness Without Reducing Accuracy
	1 Introduction
	2 Related Works
		2.1 Data Augmentation
		2.2 Adversarial Training
	3 Aligned Mixup and Adversarial Training
		3.1 Preliminaries
		3.2 Proposed Training Method
	4 Experiments
		4.1 Robustness Evaluation
		4.2 Performance on WideResNet
		4.3 Performance to the Stronger Attack
		4.4 Analysis
	5 Conclusion
	References
Multi-party Secure Comparison of Strings Based on Outsourced Computation
	1 Introduction
	2 Related Work
	3 Preliminary
		3.1 String Comparison and Hamming Distance
		3.2 Goldwasser-Micali Cryptosystem
		3.3 Secure Outsourcing Computation
		3.4 Security Model
	4 Outsourcing Secure Computation Scheme
		4.1 Efficient String Comparison
		4.2 Security Analysis
		4.3 Application Scenario: Gene Data Comparison
	5 Experiments Analysis
		5.1 Data Sources
		5.2 Data Encoding
		5.3 Test Environment
		5.4 Performance
	6 Conclusion
	A Appendix
	References
Highway: A Super Pipelined Parallel BFT Consensus Algorithm for Permissioned Blockchain
	1 Introduction
	2 System Model
	3 Highway
		3.1 Preliminaries
		3.2 The Protocol
		3.3 Decentralize Reconfiguration
	4 Safety and Liveness
		4.1 Safety
		4.2 Liveness
	5 Evaluation
	6 Conclusion
	References
Overview of DDoS Attack Research Under SDN
	1 Introduction
		1.1 Overview of SDN Architecture
		1.2 Introduction to DDoS Attacks
	2 DDoS Attack Detection Based on Mathematical Statistics
		2.1 Detection Methods Based on Entropy
		2.2 Detection Methods Based on Flow Pattern Analysis
	3 DDoS Attack Detection Based on Machine Learning
	4 DDoS Attack Detection Based on New Architecture
	5 Conclusion and Future Work
	References
A Medical Image Segmentation Method Based on Residual Network and Channel Attention Mechanism
	1 Introduction
	2 Proposed Method
		2.1 Network Structure
		2.2 Squeeze and Excitation Block
		2.3 Residual Network
	3 Experiments
		3.1 Comparison with Other Methods
		3.2 Ablation Study
	4 Conclusion
	References
Performance Improvement of Classification Model Based on Adversarial Sample Generation
	1 Introduction
	2 Background
		2.1 Adversarial Sample Attack Method
		2.2 The Similarity Evaluation Index
	3 Model Design
		3.1 Word Segmentation
		3.2 Semantic Vector Extraction
		3.3 Sorting by Importance of Words
	4 Mutation Strategies
		4.1 Single Mutation
		4.2 Combination of Mutations
	5 Model Retraining
	6 Experiments
	7 Result
		7.1 Result of Single Mutation
		7.2 Result of Combinatorial Mutations
	8 Conclusion
	References
Research on Detection Method of Large-Scale Network Internal Attack Based on Machine Learning
	1 Introduction
	2 Construction of Large-Scale Network Internal Attack Detection Model
	3 Large Scale Network Internal Attack Detection Based on Machine Learning
		3.1 Training Based on Machine Learning Model
		3.2 Attack Injection Point Analysis
		3.3 Network Internal Attack Detection Process
	4 Experiment
		4.1 Examples of Experimental Vulnerabilities
		4.2 Experimental Parameters and Environment
		4.3 Experimental Index
		4.4 Experimental Results and Analysis
	5 Conclusion
	References
Federated Community Detection in Social Networks
	1 Introduction
	2 Related Work
	3 Preliminary
		3.1 Louvain Algorithm
		3.2 Two Trapdoors Public-Key Cryptosystem
	4 Problem Formulation
		4.1 System Model
		4.2 Attacker Model
		4.3 Design Goals
	5 FedCD Framework
		5.1 Secure FPN Argmax
		5.2 Overall Framework Design
		5.3 Secure Community Construction
	6 Security Analysis
	7 Experiments Analysis
		7.1 Performance of SFAM Protocol
		7.2 Performance of FedCD Framework
	8 Conclusion
	References
A Textual Adversarial Attack Scheme for Domain-Specific Models
	1 Introduction
	2 Related Work
	3 Methodology
		3.1 Domain-Specific Perturbation Space Augmentation
		3.2 Beam Search Based Synonym Substitution
		3.3 Complete Process of Adversarial Example Generation
	4 Experiments
		4.1 Experimental Settings
		4.2 Attack Performance Results and Analysis
	5 Conclusion
	References
An Improved Conv-LSTM Method for Gear Fault Detection
	1 Introduction
	2 Related Research
	3 Methodology
		3.1 Convolutional Neural Networks
		3.2 Long Short-Term Memory Neural Networks
		3.3 Improved Conv-LSTM Fault Diagnosis Model
	4 Experimental Results and Analysis
		4.1 Experimental Setup
		4.2 Experiments Based on the Test Rig
	5 Conclusion
	References
Extracting Random Secret Key Scheme for One-Time Pad Under Intelligent Connected Vehicle
	1 Introduction
	2 On-Board Power Supply
	3 Overview of Our Approach
	4 Filtering
	5 Extracting a Perfect Random Key
		5.1 Markov Modeling
		5.2 Randomness Extractor
		5.3 Randomness Testing
		5.4 Section Summary
	6 Conclusion
	References
Semi-supervised Learning with Nearest-Neighbor Label and Consistency Regularization
	1 Introduction
	2 Related Work
		2.1 Self-training
		2.2 Consistency Regularization
		2.3 Nearest Neighbors Search in Computer Vision
	3 Method
		3.1 Overview
		3.2 Augmentation
		3.3 Support Set
		3.4 Loss Functions
	4 Experiment
		4.1 Implement Details
		4.2 Main Results
		4.3 Ablation Study
		4.4 Visualization
	5 Conclusion
	References
Bipolar Picture Fuzzy Graph Based Multiple Attribute Decision Making Approach–Part I
	1 Introduction
	2 Preliminary
	3 Bipolar Picture Fuzzy Graph-Based MADM Algorithm
	4 Numerical Example for Algorithm A
	5 Conclusion
	References
Priv-IDS: A Privacy Protection and Intrusion Detection Framework for In-Vehicle Network
	1 Introduction
		1.1 Background
	2 Preliminaries
		2.1 DP
		2.2 LDP
		2.3 Binary Random Response (BRR) Mechanism
	3 Priv-IDS Model
		3.1 -LDP Method
		3.2 Priv-IDS Model
	4 Experiments and Evaluation
		4.1 Dataset
		4.2 Evaluation Metrics
		4.3 The Parameters of -LDP
		4.4 Priv-IDS Performance
		4.5 Comparison
	5 Conclusion
	References
Dynamic Momentum for Deep Learning with Differential Privacy
	1 Introduction
	2 Releted Work
		2.1 Optimize the DP-SGD Algorithm
		2.2 Optimize Training Parameters
		2.3 Adaptive Setting Momentum
	3 Preliminaries
		3.1 Differential Privacy
		3.2 Differentially Private Stochastic Gradient Descent
		3.3 Momentum SGD
	4 Approach
		4.1 DP-SGD with MomentumSettings for the Experiments in This Section Can Be Found in Section5.
		4.2 Dynamic Momentum of DP-SGD
	5 Experiment
		5.1 Datasets
		5.2 Network and Hyperparameters
		5.3 Hardware and Software
		5.4 Experimental Results
	6 Conclusion
	A DP-SGD
	References
An Unsupervised Surface Anomaly Detection Method Based on Attention and ASPP
	1 Introduction
	2 Related Work
		2.1 Reconstruction-Based Methods
		2.2 Pre-trained Network-Based Methods
	3 Proposed Approach
		3.1 Network Architecture
		3.2 SE Block
		3.3 Atrous Spatial Pyramid Pooling
		3.4 Training Objective
	4 Experiments
		4.1 Dataset
		4.2 Evaluation Metrics
		4.3 Implementation Details
		4.4 Comparison with Other Methods
		4.5 Ablation Studies
	5 Conclusion
	References
PCB Defect Detection Method Based on Improved RetinaNet
	1 Introduction
	2 Related Work
		2.1 Attentional Mechanism
	3 Method
		3.1 Improved Backbone
		3.2 ResNet-D Residual Structure
		3.3 Channel Attention-ECA Module
		3.4 Cosine Annealing Scheduling Method
	4 Experimental Results
		4.1 Data and Parameter Settings
		4.2 Comparative Experiments and Analysis of Results
	5 Conclusion
	References
A Method of Protecting Sensitive Information in Intangible Cultural Heritage Communication Network Based on Machine Learning
	1 Introduction
	2 Intangible Cultural Heritage Target Tracking Based on Machine Learning
		2.1 Distance Value Measurement
		2.2 Random Measurement Matrix
		2.3 Compression Characteristics
	3 Protection of Sensitive Information of Intangible Cultural Heritage Communication Network (Based on Data)
		3.1 Encryption of Sensitive Information
		3.2 OSBE Protocol
		3.3 Sensitive Information Processing Flow
	4 Case Analysis
	5 Conclusion
	References
Decision Making Analysis of Traffic Accidents on Mountain Roads in Yunnan Province
	1 Introduction
	2 Preliminary
		2.1 Bipolar Fuzzy Graph
		2.2 Road Structure in Yunnan Province
	3 A Case Study on Traffic Accident on Mountain Roads in Yunnan Province
		3.1 Setting of Bipolar Fuzzy Graph
		3.2 Algorithm Description
		3.3 Aggregation Operator
		3.4 Result and Analysis
	4 Conclusion
	References
Deep Adaptively Feature Extracting Network for Cervical Squamous Lesion Cell Detection
	1 Introduction
	2 Related Work
		2.1 Medical Image Analysis
		2.2 Deep Learning Based Cervical Cell Recognition
		2.3 Cervical Cell Detection Model
	3 Method
	4 Experiments and Results
		4.1 Dataset
		4.2 Baseline Methods
		4.3 Evaluation Metrics
		4.4 Implementation Details
		4.5 Ablation Studies
		4.6 Comparison Studies
		4.7 Transfer Learning Studies
	5 Conclusion
	References
DSGRAE: Deep Sparse Graph Regularized Autoencoder for Anomaly Detection
	1 Introduction
	2 Preliminary Knowledge
	3 The Proposed Method
		3.1 Graph Regularization Term
		3.2 Sparse Constraint Term
		3.3 Deep Sparse Graph Regularized Autoencoder
		3.4 Anomaly Detection
	4 Experiments
		4.1 Benchmark Datasets
		4.2 Evaluation Criteria
		4.3 Experimental Results
	5 Conclusions
	References
A Lattice-Based Aggregate Signature Based on Revocable Identity
	1 Introduction
	2 Preliminary
		2.1 Lattice
		2.2 RSIS
		2.3 Basic Algorithm in Lattice Signature
		2.4 KUNode Algorithm.
	3 Lattice-Based Aggregated Signature Scheme Based on Revocable Identity
		3.1 Revocable Identity-Based Aggregate Signature Model
		3.2 Our Aggregate Signature Scheme
	4 Performance Analysis
		4.1 Correctness
		4.2 Efficiency Analysis
	5 Conclusion
	References
Research and Design of an Emergency Supply Assurance Monitoring System in the Post-epidemic Context
	1 Introduction
		1.1 Research Background and Purpose
		1.2 The Main Problems that Need to Be Solved at Present
		1.3 Research Direction and Goals
		1.4 The Main Content of the Research
		1.5 Research and Solve from the Following Three Key Issues
		1.6 Project Features and Innovations
		1.7 System Implementation Plan
		1.8 Research Technology Route
		1.9 Application Prospects and Social Benefits
	References
Face Presentation Attack Detection Based on Texture Gradient Enhancement and Multi-scale Fusion
	1 Introduction
	2 Related Work
		2.1 Handcrafted Feature-Based Methods
		2.2 Deep Learning Based Methods
	3 Preliminaries
		3.1 ConvNeXt Network Structure
		3.2 Sobel Operator
		3.3 Depthwise Separable Convolution
		3.4 Dilated Convolution
	4 The Proposed Scheme
		4.1 Architecture of FPAD-TGE
		4.2 Depth Separable Residual Gradient Enhancement Block
		4.3 Multi-scale Dilated Convolution Block
	5 Experiment and Analysis
		5.1 Experiment Preparation
		5.2 Experimental Results and Analysis
	6 Conclusions
	References
Optimal Revenue Analysis of the Stubborn Mining Based on Markov Decision Process
	1 Introduction
	2 Related Work
	3 Blockchain Mining Model
		3.1 State Space
		3.2 Action Space
		3.3 Transition and Reward Matrices
		3.4 Objective Function
	4 Experimental Evaluation
	5 Conclusion
	References
Bipolar Picture Fuzzy Graph Based Multiple Attribute Decision Making Approach-Part II
	1 Introduction
	2 A Case Study
	References
Machine Learning Based Method for Quantifying the Security Situation of Wireless Data Networks
	1 Introduction
	2 Research on Quantitative Assessment of Network Security Situation
		2.1 Selection of Network Security Situation Assessment Indicators
		2.2 Steps of Quantitative Assessment of Network Security Situation
		2.3 Wireless Data Network Time Series Acquisition
		2.4 Implementation of Quantitative Assessment of Cyber Security Situation
	3 Experiment and Result Analysis
		3.1 Preparation for Experiment
		3.2 Analysis of Experimental Results
	4 Conclusions
	References
Overlapping Community Discovery Algorithm Based on Three-Level Neighbor Node Influence
	1 Introduction
	2 OCDITN Algorithm
		2.1 Influence Measurement Based on Three Level Neighbors
		2.2 Label Updating of OCDITN Algorithm
		2.3 OCDITN Algorithm
	3 Experiment and Analysis
		3.1 Experimental Results and Analysis of Real Data Sets
		3.2 Experimental Results and Analysis of Simulated Data Sets
	4 Conclusion
	References
Content-Aware Deep Feature Matching
	1 Introduction
	2 Method
		2.1 Pre-alignment
		2.2 Content-Aware Learning
	3 Experiments
	4 Conclusions
	References
F2DLNet: A Face Forgery Detection and Localization Network Based on SSIM Error Maps
	1 Introduction
	2 Related Work
		2.1 Face Forgery Detection
		2.2 Face Manipulation Localization
	3 Generation of Localization Map
		3.1 Error-Threshold Method
		3.2 Localization Map Generation Based on SSIM
	4 The Proposed F2DLNet
		4.1 Network Architecture
		4.2 Adjacent Layer Aggregation Module
		4.3 Gradient-Enhanced Block
		4.4 Loss Functions
	5 Experimental Results and Analysis
		5.1 Implementation Details
		5.2 Datasets and Evaluation Metrics
		5.3 Experiments and Performance Analysis
	6 Conclusion
	References
An Eye-Gaze Tracking Method Based on a 3D Ocular Surface Fitting Model
	1 Introduction
	2 Eye Point Cloud Data Acquisition Based on Binocular Stereo Vision
	3 Acquisition of Gaze Direction Based on 3D Ocular Surface Fitting
		3.1 Ocular Geometrical and Optical Structure Analysis
		3.2 Ocular Surface Fitting and Feature Extraction
		3.3 Estimation of Optical Axis and Acquisition of Gaze Direction
	4 Experiments of Gaze Tracking and Results Analysis
		4.1 Experimental Devices and Parameters Acquisition
		4.2 Establishment of 3D Coordinate System
		4.3 Experimental Results and Analysis
	5 Conclusions
	References
A Certificateless-Based Blind Signature Scheme with Message Recovery
	1 Introduction
	2 Preliminaries
		2.1 Billinear Pairing
		2.2 Framework of an Efficient CL-BS-MR
	3 A Certificateless-Based Blind Signature Scheme with Message Recovery
	4 Analysis of Proposed CL-BS-MR Scheme
		4.1 Correctness Analysis
		4.2 Security Analysis
	5 Conclusion
	References
Fault Detection of Rolling Bearings by Using a Combination Network Model
	1 Introduction
	2 Methodology
		2.1 Convolutional Neural Network
		2.2 Random Forest
		2.3 Combined Network Model
	3 Experiments
		3.1 Data Sources
		3.2 Experiments and Results
	4 Conclusion
	References
zkChain: An Efficient Blockchain Privacy Protection Scheme Based on zk-SNARKs
	1 Introduction
		1.1 Related Work
		1.2 Our Contributions
	2 Preliminaries
	3 Architecture
		3.1 zk-SNARK Scheme
		3.2 Privacy Protection Scheme
	4 Proposed zkChain Scheme
		4.1 Mint
		4.2 Verify
		4.3 Transfer
		4.4 Receive
	5 Experiments
		5.1 Computing and Storage Performance of zkSNARK
		5.2 Processing Performance Based on FISCO BCOS
	6 Conclusion
	References
Research on Influential Factors of Online Learning Behavior Based on Big Data
	1 Introduction
	2 Research Review
	3 Analysis of Online Learning Behavior Status Quo
		3.1 Value Analysis of Online Learning Data
		3.2 Takeholder Analysis of Online Learning
	4 Construction of Influencing Factors Model of Online Learning Behavior in the Big Data
		4.1 Construction of Online Learning Learner's Own Literacy Factor Model
		4.2 Modeling of External Influencing Factors Related to Online Learning Behavior
	5 Conclusion
	References
Short Speech Key Generation Technology Based on Deep Learning
	1 Introduction
	2 Related Work
	3 Voiceprint Key Generation Model
	4 Experiment
	5 Conclusion and Future Work
	References
Domain Adversarial Interaction Network for Cross-Domain Fault Diagnosis
	1 Introduction
	2 Methodologies
		2.1 Problem Statement
		2.2 Feature Extractor
		2.3 Fault Classifier
		2.4 Domain Discriminator
		2.5 Feature Distribution Alignment
		2.6 Objective Funtion
	3 Experiments
		3.1 Data Description
		3.2 Compared Approaches
		3.3 Experimental Setting
		3.4 Experimental Results
	4 Conclusion
	References
A Vehicle Data Publishing System with Privacy-Awares in VANETs Based on Blockchain
	1 Introduction
	2 Preliminary
		2.1 Bilinear Pairings
		2.2 DLP Hard Problem
	3 A Vehicle Data Publishing System with Privacy-Awares in VANETs Based on Blockchain
		3.1 The Entities in the System
		3.2 Design Goals
		3.3 System Initialization
		3.4 Publishing Data
		3.5 Tracking Vehicle
	4 Performance Analysis
		4.1 Correctness Proof
		4.2 Efficiency Analysis
	5 Conclusions
	References
Highsimb: A Concrete Blockchain High Simulation with Contract Vulnerability Detection for Ethereum and Hyperledger Fabric*-12pt
	1 Introduction
	2 Related Works
	3 Our Concrete Blockchain High Simulation Model
		3.1 Our Blockchain Abstraction Model
		3.2 Our High Simulation Platform Semantic Metamodel
	4 Our Concrete Blockchain High Simulation Empirical Construction
		4.1 Concrete Design Methodology
		4.2 Personalized Blockchain Model Simulation Strategy
	5 Conclusion and Future Works
	References
Research on Key Technologies for the Trusted Perception of Network Information for Big Data
	1 Introduction
	2 Analysis and Research of Key Technical Models
		2.1 Analysis and Comparison of Mainstream Models
		2.2 Multi-dimensional Analysis of Intelligence Source Credibility Assessment Method
	3 Specific Analysis of Situational Awareness
		3.1 Situational Awareness Method Proposed
		3.2 Situational Awareness Method Analysis
		3.3 Results and Analysis
	4 Conclusion
	References
Micro-expression Recognition Method Combining Dual-Stream Convolution and Capsule Network
	1 Introduction
	2 Proposed Methods
		2.1 Preprocessing
		2.2 Dual-Stream Convolutional Feature Extraction Network
		2.3 Capsule Network Layer
	3 Experiments
		3.1 Dataset and Experimental Evaluation Metrics
		3.2 Experimental Setup and Preprocessing
		3.3 Experimental Results and Analysis
	4 Conclusion
	References
Security Scheduling Method of Cloud Network Big Data Cluster Based on Association Rule Algorithm
	1 Introduction
	2 Analysis of Cloud Network Security Performance Based on Association Rule Algorithm
		2.1 Association Rule Mining Steps
		2.2 Hadoop/Mapreduce Framework
		2.3 Security Mapping Conditions
	3 Security Scheduling and Processing of Big Data Cluster
		3.1 Link Bandwidth - Time List
		3.2 SDN Dispatching System
		3.3 Packet Routing Coefficient
	4 Example Analysis
	5 Conclusion
	References
Towards Differentially Private Contrastive Learning
	1 Introduction
	2 Background and Related Work
		2.1 Contrastive Learning
		2.2 Differential Privacy
	3 Differentially Private Contrastive Learning
		3.1 Basic Components of Private Training
		3.2 Private Training of SimCLR
	4 Privacy Analysis
		4.1 Privacy Analysis of SimCLR
	5 Implementation
		5.1 Experimental Setup
		5.2 Experimental Results
	6 Conclusion
	References
Two-Stage High Precision Membership Inference Attack
	1 Introduction
	2 Background
		2.1 Membership Inference Attacks
		2.2 Calibrated Membership Inference Attacks
	3 Methodology
		3.1 Threat Model
		3.2 Two-stage HP-MIA
	4 Experiments
		4.1 Experimental Setup
		4.2 Attack Evaluation
		4.3 Ablation Experiments
	5 Conclusion
	A  Algorithm for the Selection of Attack Thresholds
	References
Secure Storage Method for Network Resources of Professional Works Based on Decision Tree Algorithm
	1 Introduction
	2 Research on Secure Storage of Network Resources of Professional Works in Design
		2.1 Building a Secure Storage System for Network Resources
		2.2 Secure Storage Network Communication Design
		2.3 Caching Model Construction Based on Decision Tree Algorithm
		2.4 Implementation of Secure Storage of Network Resources
	3 Experiment and Result Analysis
		3.1 Preparation for Experiment
		3.2 Analysis of Experimental Results
	4 Conclusions
	References
Vehicle CAN Network Intrusion Detection Model Based on Extreme Learning Machine and Feature Analysis
	1 Introduction
	2 Design of Intrusion Detection Model for Automotive Can Network
		2.1 Simulation of Car CAN Network Intrusion Behavior
		2.2 Construction of Automotive CAN Network Structure
		2.3 Collect and Process Automotive CAN Network Traffic Data
		2.4 Using Extreme Learning Machine to Analyze Network Traffic Characteristics
		2.5 Realizing Intrusion Detection of Automotive CAN Network
	3 Performance Test Experimental Analysis
		3.1 Configuring the Automotive CAN Network
		3.2 Prepare Experimental Data Samples
		3.3 Set the Operating Parameters of the Extreme Learning Machine Algorithm
		3.4 Describe the Performance Test Experiment Process
		3.5 Analysis of Results
	4 Conclusion
	References
A Broad Learning System Based on the Idea of Vertical Federated Learning
	1 Introduction
	2 Research Background
		2.1 Vertical Federated Machine Learning
		2.2 Broad Learning System
	3 Design Goals and System Models
		3.1 Design Goal
		3.2 System Model
	4 Analysis of Broad Learning System Based on Vertical Federation Idea
		4.1 Security Analysis
		4.2 Analysis of the Tasks Undertaken by Each Part
		4.3 Inadequacies of the Proposed System
	5 Experimental Results
		5.1 Experimental Process
		5.2 Experimental Results and Analysis
	6 Conclusion
	References
PAMP: A New Atomic Multi-Path Payments Method with Higher Routing Efficiency
	1 Introduction
		1.1 Related Works
		1.2 Our Contributions
	2 Preliminary
		2.1 How to Realize Off-Chain Payment?
		2.2 How to Find an Optimal Transaction Payment Route?
	3 PAMP Design
	4 Evaluation
	5 Summary and Future Work
	References
Privacy-Preserving Searchable Encryption Scheme Based on Deep Structured Semantic Model over Cloud Application
	1 Introduction
	2 Related Works
	3 Problem Formulation
		3.1 Notations
		3.2 System Model
		3.3 Threat Model
		3.4 Design Goal
	4 Framework for DSRSE and Structure of Neural Network
		4.1 Framework
		4.2 Deep Convolutional Neural Networks Model
		4.3 Distributed Convolutional Neural Network Training
	5 Proposed Scheme
		5.1 DSRSE Scheme
		5.2 Building the Document Index Clustering Tree
		5.3 Searching in the Document Index Clustering Tree
		5.4 Updating the Document Index Clustering Tree
	6 Security Analysis
		6.1 Documents Privacy
		6.2 Privacy of Document Indexes and Trapdoors
	7 Performance Evaluation
		7.1 Trapdoor Construction Efficiency
		7.2 Document Retrieval Accuracy
		7.3 Document Retrieval Efficiency
		7.4 Communication Overhead
	8 Conclusion
	A Appendix
	References
A Event Extraction Method of Document-Level Based on the Self-attention Mechanism
	1 Introduction
		1.1 A Subsection Sample
	2 Related Work
	3 Methods
		3.1 Sentence Level Feature Acquisition
		3.2 Sentence Package Strategy
		3.3 Document-Level Feature Acquisition
		3.4 Classifier
	4 Experiment
		4.1 Datasets
		4.2 Evaluation Indicators
		4.3 Parameter Settings
		4.4 Experimental Results
	5 Conclusion
	References
Author Index