
Data Science
Description
Alles über E-Books | Antworten auf Fragen rund um E-Books, Kopierschutz und Dateiformate finden Sie in unserem Info- & Hilfebereich.
This two volume set (CCIS 1628 and 1629) constitutes the refereed proceedings of the 8th International Conference of Pioneering Computer Scientists, Engineers and Educators, ICPCSEE 2022 held in Chengdu, China, in August, 2022.
The 65 full papers and 26 short papers presented in these two volumes were carefully reviewed and selected from 261 submissions. The papers are organized in topical sections on: Big Data Mining and Knowledge Management; Machine Learning for Data Science; Multimedia Data Management and Analysis.
More details
Other editions
Additional editions

Content
- Intro
- Preface
- Organization
- Contents - Part I
- Contents - Part II
- Big Data Mining and Knowledge Management
- Self-attention Based Multimodule Fusion Graph Convolution Network for Traffic Flow Prediction
- 1 Introduction
- 2 Spatiotemporal Prediction in Deep Learning
- 2.1 Time Correlation Research
- 2.2 Time Correlation Research
- 3 Prediction Model of Traffic Flow Based on Multi-module Fusion
- 3.1 Model Frame Diagram
- 3.2 Space-Time Decoupling
- 3.3 Spatial Convolution
- 3.4 Spatial Self-attention
- 3.5 Temporal Convolution
- 3.6 Time Self-attention
- 3.7 Information Fusion and GRU
- 4 Experimental Analysis
- 4.1 Dataset
- 4.2 Analysis of Results
- 5 Conclusion
- References
- Data Analyses and Parallel Optimization of the Tropical-Cyclone Coupled Numerical Model
- 1 Introduction
- 1.1 A Subsection Sample
- 2 Model Setup
- 2.1 Atmospheric Model Setup
- 2.2 Hydrodynamic Model Setup
- 2.3 Ocean Wave Model Setup
- 2.4 HPC Facilities
- 2.5 Coupled Variables
- 3 Scaling Experiments
- 3.1 Parallel Tests Analysis
- 3.2 SWAN Model Parallel Algorithm Optimization
- 3.3 Ocean Model Grid Optimization
- 4 Parallel Test Results
- 5 Model Results Discussion
- 6 Conclusion
- References
- Factorization Machine Based on Bitwise Feature Importance for CTR Prediction
- 1 Introduction
- 2 Related Work
- 3 Our Approach
- 3.1 Embedding Layer
- 3.2 Learning
- 4 Experiments
- 4.1 Experimental Settings
- 4.2 Hyperparameter Study
- 4.3 Ablation Study
- 4.4 Performance Comparison
- 5 Conclusion
- References
- Focusing on the Importance of Features for CTR Prediction
- 1 Introduction
- 2 ECABiNet Model
- 2.1 Sparse Input and Embedding Layer
- 2.2 Layer Norm
- 2.3 ECANET Layer
- 2.4 Feature Cross Layer
- 2.5 DNN Layer
- 2.6 Output
- 3 Experiment
- 3.1 Experimental Setup
- 3.2 LayerNorm Effect Comparison
- 3.3 Comparison of the Effects of Different Attention Modules
- 3.4 Comparison of the Classic Model
- 3.5 Study HyperParameter
- 4 Related Work
- 5 Conclusions
- References
- Active Anomaly Detection Technology Based on Ensemble Learning
- 1 Introduction
- 2 Problem Statement
- 3 Proposed Model
- 3.1 Supervised Ensemble Learning Model
- 3.2 Human Participation
- 3.3 Model Self-training
- 3.4 Experiment
- 3.5 Conclusion
- References
- Automatic Generation of Graduation Thesis Comments Based on Multilevel Analysis
- 1 Introduction
- 2 Technical Principle
- 2.1 BERT Model Introduced
- 2.2 Basic Structure of the BERT Model
- 2.3 Comparison with Other Algorithms
- 3 Project Analysis
- 3.1 Technical Route
- 3.2 Technical Analysis
- 4 Project Implementation
- 4.1 Database Established Modification
- 4.2 Student Information Input
- 4.3 Neural Network Training
- 4.4 Automatically Generate Comments
- 5 Conclusion
- References
- A Survey of Malware Classification Methods Based on Data Flow Graph
- 1 Introduction
- 2 Data Flow Graph
- 2.1 Basic Concepts of the Data Flow Graph
- 2.2 Data Flow Graph Corresponding to Common APIs
- 2.3 Extension of Data Flow Graphs
- 3 Malware Classification Based on Data Flow Graph
- 3.1 User-Defined Data Flow Graph Feature-based Malware Classification
- 3.2 Data Flow Graph Similarity-Based Malware Classification
- 3.3 Graph Neural Network-Based Malware Classification
- 4 Discussion
- 5 Conclusion
- References
- Anomaly Detection of Multivariate Time Series Based on Metric Learning
- 1 Introduction
- 2 Preliminaries
- 3 Proposed Model
- 3.1 Preprocessing
- 3.2 Encoder for High-Dimensional Time Series Data
- 3.3 Attentional Center Learning
- 3.4 Loss Function
- 3.5 Semisupervised Learning
- 4 Experiments
- 4.1 Dataset
- 4.2 Setup
- 4.3 Result
- 5 Conclusion
- References
- Social Network Analysis of Coauthor Networks in Inclusive Finance in China
- 1 Introduction and Motivation
- 2 Data Collection and Preprocessing
- 3 Results
- 3.1 General Characteristics of the Coauthor Network
- 3.2 Ego Characteristics of the Coauthor Network
- 3.3 The Evolution of Cohesive Subgroups in the Coauthor Network
- 4 Conclusions
- References
- Multirelationship Aware Personalized Recommendation Model
- 1 Introduction
- 2 Preliminary Preparation
- 2.1 Problem Definition
- 2.2 Data Preprocessing
- 2.3 User Relationship Graphs
- 3 Modeling and Training
- 3.1 MrAPR Model
- 3.2 Model Training
- 4 Experiment
- 4.1 Dataset
- 4.2 Baselines and Evaluation Metrics
- 4.3 Parameter Settings
- 4.4 Ablation Experiments
- 5 Conclusion
- References
- Machine Learning for Data Science
- Preliminary Study on Adapting ProtoPNet to Few-Shot Learning Using MAML
- 1 Introduction
- 2 Related Work
- 2.1 Few-Shot Learning
- 2.2 Interpretability
- 3 Proposed Methods
- 3.1 Adapting ProtoPNet to MAML
- 3.2 Evaluating Models
- 4 Experiments
- 4.1 Datasets
- 4.2 Experiment 1: Omniglot Few-Shot Classification
- 4.3 Experiment 2: MiniImagenet Few-Shot Classification
- 4.4 Experiment 3: Interpretability Analysis on Omngilot
- 4.5 Experiment 4: Preliminary Interpretability Analysis on MiniImagenet
- 5 Conclusion and Future Work
- References
- A Preliminary Study of Interpreting CNNs Using Soft Decision Trees
- 1 Introduction
- 2 Related Work
- 3 Proposed Methods
- 3.1 Model Foundations
- 3.2 Using Normal/Soft Decision Trees to Interpret CNNs
- 3.3 Evaluating Interpretability
- 4 Experiments
- 4.1 Dataset and Experimental Setup
- 4.2 Experiment 1: Classification Performance
- 4.3 Experiment 2: Visualization of Normal/Soft Decision Trees' Top Features
- 4.4 Experiment 3: Interpretability Performance
- 4.5 Experiment 4: Scores of Human Experts on Tag Clarity
- 5 Conclusion and Future Work
- References
- Deep Reinforcement Learning with Fuse Adaptive Weighted Demonstration Data
- 1 Introduction
- 2 Related Work
- 2.1 Deep Reinforcement Learning
- 2.2 Multiagent Reinforcement Learning
- 3 Methods
- 4 Experimental Results and Analysis
- 4.1 Experimental Environment and Data
- 4.2 Results and Analysis
- 5 Discussion
- References
- DRIB: Interpreting DNN with Dynamic Reasoning and Information Bottleneck
- 1 Introduction
- 2 Related Works
- 2.1 Explain the Existing Deep Learning Models
- 2.2 Construction of Interpretable Deep Learning Models
- 3 Method
- 3.1 Dynamic Reasoning Decision Module
- 3.2 Information Bottleneck Verification Module
- 4 Experiments
- 4.1 Experimental Settings
- 4.2 Interpretability of Calculation in Dynamic Reasoning Decision
- 4.3 Explainability of Attribution in the Information Bottleneck
- 4.4 Visualization of Understandability
- 5 Conclusion
- References
- Multimedia Data Management and Analysis
- Advanced Generative Adversarial Network for Image Superresolution
- 1 Introduction
- 2 Related Work
- 3 GAN and SRGAN
- 4 Proposed Method
- 4.1 Generator Network Structure
- 4.2 Discriminator Network Structure
- 4.3 Loss Function
- 5 Experiment Results and Analysis
- 5.1 Implementation Details
- 5.2 Datasets and Evaluation Metrics
- 5.3 Experimental Results and Analysis
- 5.4 Ablation Study
- 6 Conclusions
- References
- Real-World Superresolution by Using Deep Degradation Learning
- 1 Introduction
- 2 Related Work
- 2.1 Real-World Superresolution
- 2.2 Contrastive Learning
- 3 PurPosed Method
- 3.1 Overview of the Unsupervised Framework
- 3.2 Degradation Model
- 3.3 Reconstruction Model
- 4 Experiments
- 4.1 Training Data
- 4.2 Training Details
- 4.3 Training Details
- 5 Conclusion
- References
- Probability Loop Closure Detection with Fisher Kernel Framework for Visual SLAM
- 1 Introduction
- 2 Related Works
- 3 Methodology
- 3.1 Fisher Vector Generation
- 3.2 Probability Visual Vocabulary
- 3.3 Loop Closure Detection
- 4 Results and Discussion
- 4.1 Dataset and Preprocessing
- 4.2 Evaluation Metrics
- 4.3 2D Motion
- 4.4 3D Motion
- 4.5 Bidirectional Loops
- 4.6 Ablation Study
- 5 Conclusions
- References
- A Complex Background Image Registration Method Based on the Optical Flow Field Algorithm
- 1 Introduction
- 2 The Proposed Method
- 3 Evaluation Functions
- 4 Experimental Results
- 5 Conclusion
- References
- Collaborative Learning Method for Natural Image Captioning
- 1 Introduction
- 2 Related Work
- 3 Method
- 3.1 P2PM: Pix2Pix Inverting Module
- 3.2 NLGM: Natural Language Generation Module
- 3.3 Collaborative Learning Framework
- 4 Experiments
- 4.1 Experimental Setup
- 4.2 Main Results
- 5 Conclusion
- References
- Visual Analysis of the National Characteristics of the COVID-19 Vaccine Based on Knowledge Graph
- 1 Introduction
- 2 Related Studies
- 3 Construction of the COVID-19 Vaccine Knowledge Graph
- 3.1 Data Acquisition
- 3.2 Entity Extraction to Construct a Relational Model
- 3.3 Knowledge Graph Establishment
- 4 Visual Analysis of the National Characteristics of the COVID-19 Vaccine
- 5 Conclusions and Recommendations
- References
- Speech Recognition for Parkinson's Disease Based on Improved Genetic Algorithm and Data Enhancement Technology
- 1 Introduction
- 2 The Proposed Methods
- 2.1 Method
- 3 Speech Recognition and Diagnosis
- 3.1 Data Preprocessing
- 3.2 Improved GA-SVM Model
- 3.3 Speech Recognition Algorithm
- 4 Experiment and Evaluation
- 4.1 Experiment Setup
- 4.2 Comparison and Analysis of Results
- 5 Conclusion
- References
- A Review of Animal Individual Recognition Based on Computer Vision
- 1 Introduction
- 2 Construction of Animal Datasets
- 2.1 Data Collection Method
- 2.2 Datasets Related to Individual Animal Identification
- 2.3 Data Preprocessing
- 3 Individual Animal Target Detection
- 3.1 Traditional Target Detection Algorithms
- 3.2 Animal Target Detection Algorithm Based on Deep Learning
- 4 Methods of Individual Identification of Animals
- 4.1 Typical Animal Individual Recognition Algorithms
- 4.2 Performance Comparison
- 5 Application of Animal Individual Identification Technology
- 6 Summary and Outlook
- 6.1 Major Problems
- 6.2 Outlook
- References
- CIRS: A Confidence Interval Radius Slope Method for Time Series Points Based on Unsupervised Learning
- 1 Introduction
- 2 Related Work
- 3 Definition
- 3.1 Multidimensional Time Series
- 3.2 Definition of Research Problem
- 4 Model
- 5 Algorithm Complexity and Feasibility Analysis
- 6 Experiment
- 6.1 Data Set
- 6.2 Set up
- 6.3 Accuracy
- 7 Conclusion
- References
- LSN: Long-Term Spatio-Temporal Network for Video Recognition
- 1 Introduction
- 2 Method
- 2.1 LSTM and ConvLSTM
- 2.2 Long-Term Spatio-Temporal Network
- 3 Experiments
- 3.1 Datasets
- 3.2 Implementation Details
- 3.3 Ablation Study
- 3.4 Comparison with Representative Methods
- 4 Conclusion
- References
- A Novel Fish Counting Method Based on Multiscale and Multicolumn Convolution Group Network
- 1 Introduction
- 1.1 Related Work
- 2 The Proposed Method
- 2.1 Fish Counting Network
- 2.2 Asymmetric Convolution
- 2.3 Spatial Pyramid Structure
- 2.4 Multicolumn Convolution Group
- 2.5 Predicted Density Plot for Count
- 2.6 Loss Function
- 2.7 Dataset Construction
- 3 Experiments
- 3.1 Evaluation Metric
- 3.2 Experimental Design
- 3.3 Results and Analysis
- 3.4 Ablation Experiment
- 4 Conclusion
- References
- A Low Spectral Bias Generative Adversarial Model for Image Generation
- 1 Introduction
- 2 Related Work
- 3 The Spectral Bias of Generated Images
- 3.1 Spectral Bias in the Resize-Convolution Generator
- 3.2 Spectral Bias in the Deconvolution Generator
- 4 The Reason of Different Spectral Bias
- 5 Solutions
- 6 Conclusion
- References
- Research on Video Object Detection Methods Based on YOLO with Motion Features
- 1 Introduction
- 2 Method Construction
- 2.1 Moving Object Detection Module
- 2.2 Object of Interest Detection Module
- 2.3 Background Decision Module
- 3 Experiment and Evaluation
- 3.1 Implementation Details
- 3.2 Dataset Preparation and YOLOv4 Model Validation
- 3.3 Motion Video Object Detection and Recognition
- 4 Conclusion
- References
- TU-Net: U-shaped Structure Based on Transformers for Medical Image Segmentation
- 1 Introduction
- 2 Method
- 2.1 Algorithm Framework
- 2.2 Patch Embedding
- 2.3 Skip Connection
- 3 Experiments
- 3.1 Dataset and Implementation Details
- 3.2 Experimental Results
- 3.3 Ablation Study
- 4 Conclusion
- References
- Reducing Video Transmission Cost of the Cloud Service Provider with QoS-Guaranteed
- 1 Introduction
- 2 Related Work
- 3 System Model and Problem Formulation
- 3.1 System Model
- 3.2 Formulaic Description of the Algorithm
- 4 Algorithm Design and Description
- 4.1 The Low Delay Algorithm for the Emergency Video Clips
- 4.2 The Minimum Cost Path Algorithm for the Nonemergency Video Clips
- 5 Evaluation
- 5.1 Simulation Setup
- 5.2 Description of Key Performance Indicators
- 5.3 Simulation Results
- 6 Conclusions
- References
- Association Strategy Graph Convolutional Neural Network for Human Skeletal Behavior Recognition
- 1 Introduction
- 2 Related Work
- 3 Algorithms and Construction
- 3.1 Equations
- 3.2 RCTR-GCN Model Definition of the New Association Policy
- 3.3 The Relational Architecture of the Model
- 4 Experiment and Analysis
- 4.1 Datasets
- 4.2 Comparison of RCTR-GCN Experiments with Different Association Strategies
- 4.3 Comparison with Other Popular Algorithms
- 5 Conclusion
- References
- Unsupervised Style Control for Image Captioning
- 1 Introduction
- 2 Related Work
- 2.1 CLIP Pretrained Model
- 2.2 Stylized Image Captioning
- 2.3 Text Style Control
- 3 Approach
- 3.1 Overview
- 3.2 Image Encoder
- 3.3 Stylized Caption Decoder
- 3.4 Image Sentiment Recognition
- 3.5 Training Strategy
- 4 Experiment
- 4.1 Datasets
- 4.2 Evaluation Metrics
- 4.3 Implementation Details
- 4.4 Results
- 4.5 Ablation Studies
- 4.6 Human Evaluation
- 5 Conclusion
- References
- AM-PSPNet: Pyramid Scene Parsing Network Based on Attentional Mechanism for Image Semantic Segmentation
- 1 Introduction
- 2 AM-PSPNet
- 2.1 Efficient Channel Attention Module
- 2.2 Deep Guidance Fusion Module
- 3 Experiments and Analysis
- 3.1 Experimental Design
- 3.2 Ablation Study
- 3.3 Performance Evaluation on PASCAL VOC 2012
- 3.4 Performance Evaluation on Cityscapes
- 4 Conclusions
- References
- Author Index
System requirements
File format: PDF
Copy protection: Watermark-DRM (Digital Rights Management)
System requirements:
- Computer (Windows; MacOS X; Linux): Use the free software Adobe Reader, Adobe Digital Editions, or any other PDF viewer of your choice (see eBook Help).
- Tablet/Smartphone (Android; iOS): Install the free app Adobe Digital Editions or another reading app for eBooks, e.g., PocketBook (see eBook Help).
- E-reader: Bookeen, Kobo, Pocketbook, Sony, Tolino and many more (only limited: Kindle).
The file format PDF always displays a book page identically on any hardware. This makes PDF suitable for complex layouts such as those used in textbooks and reference books (images, tables, columns, footnotes). Unfortunately, on the small screens of e-readers or smartphones, PDFs are rather annoying, requiring too much scrolling.
This eBook uses Watermark-DRM, a „soft” copy protection. This means that there are no technical restrictions to prevent illegal distribution. However, there is a personalised watermark embedded in the eBook that can be used to identify the purchaser of the eBook in the event of misuse and to provide evidence for legal purposes.
For more information, see our eBook Help page.