
Complex Networks & Their Applications XII
Description
Alles über E-Books | Antworten auf Fragen rund um E-Books, Kopierschutz und Dateiformate finden Sie in unserem Info- & Hilfebereich.
More details
Other editions
Additional editions

Persons
Hocine Cherifi received the Ph.D. from the National Polytechnic Institute, Grenoble, in 1984. He has been a professor of computer science at the University of Burgundy, Dijon, France, since 1999. Before moving to Dijon, he held faculty positions with Rouen University and Jean Monnet University, France. He has also held visiting positions with Yonsei University, South Korea; the University of Western Australia, Australia; the National Pintung University, Taiwan; and Galatasaray University, Turkey. He has published over 200 scientific papers in international refereed journals and conference proceedings. His current research interests include computer vision and complex networks. He held leading positions in more than 15 international conference organizations as the general chair and the program chair. He has served on more than 100 program committees. He is the Founder of the International Conference on Complex Networks and their Applications. He is a member of the editorial board of Computational Social Networks, PLOS One, IEEE Access, Journal of Imaging, Complex Systems, Quality and Quantity, and Scientific Reports. He is the Founding Editor-in-Chief of Applied Network Science and PLOS Complex Systems.
Chantal CHERIFI received her PhD in Computer science from Corsica University, France, in 2011. Since 2014, she has worked as an Associate Professor at the DISP laboratory at the University of Lyon2, France. Her main research interests are information systems agility and big data management with applications on enterprise information systems and smart cities, using complex networks, ontologies, and Product Lifecycle Management (PLM) systems tools. She is involved in several International conference organizations: Complex Networks (Program Chair & Local Committee 2017, Poster Chair 2016), CompleNet 2016 (Poster Chair), PLM 2015 (Local Committee), and DICTAP 2011 (Local Committee). She serves as a member of International Conferences Program committees (Complex Networks [2017, 2016], CompletNet 2016, Complexis 2016, ISCRAM-med [2017, 2016, 2015], CSCESM 2014, SITIS [2015, 2014, 2013, 2012], ICIEIS [2013, 2011]) and Journal referee (IJCIM 2016, EPL 2015, Scientific Reports - Nature 2015) She is a member of EU Erasmus-Mundus programs (SmartLink (2012-2016), cLink program (2012-2016)). Her local responsibilities include being a Committee Lab member (Since 2016)and Lab seminar coorganizer (Since 2014).
Murat Donduran is a Professor of Economics at the Department of Economics, YILDIZ Technical University Istanbul Türkiye, where he is the Director of the Graduate School of Social Sciences. He is also the Board member of in Turkish Economic Foundation. His research is on microeconomics, computational economics, firm dynamics and game theory. He received his Bachelor's and Master's Degrees in Economics from Marmara University, Istanbul, Türkiye, and a Ph.D. in Economics in 2000 from the YILDIZ Technical University. Since 2000 he has been a faculty member in the Faculty of Economic and Administrative Sciences at the YILDIZ Technical University. He has organized several conferences in the field such as YILDIZ International Conference on Social Sciences, the Annual International Conference on Social Sciences (2016-2020), and the International Conference on Economics (IceTea) 2019-2023. He has published many articles in scientific journals such as Games and Economic Behavior, Review of Industrial Organization, Physica A, and North American Journal of Economics and Finance.Content
- Intro
- Preface
- Organization and Committees
- Contents
- Graph Neural Networks
- Network Design Through Graph Neural Networks: Identifying Challenges and Improving Performance
- 1 Introduction
- 2 Related Work
- 3 Notation
- 4 Optimization for Network Editing
- 4.1 Graph Properties that Influence Edge Scores
- 4.2 ORE: Improved Edge Editing
- 5 Experimental Setup
- 5.1 Network Editing Process
- 5.2 Learning Tasks
- 6 Results
- 6.1 Motif Detection
- 6.2 Shortest Path, Triangle Counting, and Graph Classification
- 7 Conclusion
- References
- Sparse Graph Neural Networks with Scikit-Network
- 1 Introduction
- 2 Related Work
- 3 Message Passing with Sparse Matrices
- 3.1 Message Passing Overview
- 3.2 Using Sparse Matrices in Scikit-Network
- 4 Scikit-Network GNN Design
- 5 Performance
- 5.1 Datasets
- 5.2 Performances and Running Time
- 5.3 Impact of Graph Characteristics
- 6 Conclusion
- References
- Enhancing Time Series Analysis with GNN Graph Classification Models
- 1 Introduction
- 2 Related Work
- 3 Methods
- 3.1 Transform EEG Data to Graphs
- 3.2 Cosines for Consecutive Years as Indicators
- 3.3 Graph Construction from Climate Data
- 3.4 From Cosine Similarity Matrices to Graph Construction
- 3.5 Implementing the GNN Graph Classification Model
- 4 Experiments
- 4.1 EEG Data Graph Classification
- 4.2 Climate Data Graph Classification
- 5 Conclusions
- References
- When Do We Need Graph Neural Networks for Node Classification?
- 1 Introduction
- 2 Preliminaries
- 2.1 Graph Laplacian and Affinity Matrix
- 3 Measuring the Effect of Edge Bias
- 3.1 Normalized Total Variation (NTV) and Normalized Smoothness Value (NSV) for Measuring Edge Bias
- 3.2 Hypothesis Testing for Edge Bias
- 3.3 Why NTV and NSV Work
- 4 Related Works
- 5 Conclusion
- A Details of NSV and Sample Covariance Matrix
- References
- Training Matters: Unlocking Potentials of Deeper Graph Convolutional Neural Networks
- 1 Introduction
- 2 Preliminaries
- 3 Energy Loss During Back Propagation
- 3.1 Forward Pass Analyses: Difficult and Complicated
- 3.2 Backward Pass Analyses: Identifying the Core Problem
- 4 Methods to Alleviate BP Energy Loss
- 4.1 Spectra Shift and Topology Rescaling (TR)
- 4.2 Weight Initialization
- 4.3 Normalization
- 4.4 Skip Connection
- 5 Experiments
- 5.1 Training Difficulty and Generalization
- 5.2 Finetuned Performance Boost
- 6 Conclusion
- References
- E-MIGAN: Tackling Cold-Start Challenges in Recommender Systems
- 1 Introduction
- 2 Related Work
- 3 The Proposed Framework Enhanced-MIGAN
- 3.1 Mutual-Interaction Graph Attention Network Recommender
- 3.2 Content-Based Embeddings
- 4 Experiments and Results
- 5 Conclusion
- References
- Heterophily-Based Graph Neural Network for Imbalanced Classification
- 1 Introduction
- 2 Related Work
- 3 Motivation
- 4 Methodology
- 4.1 Im-GBK
- 4.2 Fast Im-GBK
- 5 Experiments
- 5.1 Experiment Settings
- 5.2 Comparisons with Baselines (RQ1)
- 5.3 Comparison in Efficiency (RQ2)
- 5.4 Ablation Analysis (RQ3)
- 6 Conclusion
- References
- TimeGNN: Temporal Dynamic Graph Learning for Time Series Forecasting
- 1 Introduction
- 2 Related Work
- 3 Method
- 3.1 Time Series Feature Extraction
- 3.2 Graph Structure Learning
- 3.3 Graph Neural Network for Forecasting
- 3.4 Training and Inference
- 4 Experimental Evaluation
- 4.1 Datasets
- 4.2 Baselines
- 4.3 Experimental Setup
- 4.4 Results
- 5 Conclusion
- References
- UnboundAttack: Generating Unbounded Adversarial Attacks to Graph Neural Networks
- 1 Introduction
- 2 Related Work
- 3 Preliminaries
- 3.1 Graph Neural Networks
- 3.2 Adversarial Attacks
- 4 Proposed Method: UnboundAttack
- 4.1 Unbounded Adversarial Attacks
- 4.2 Architecture Overview
- 4.3 Training and Loss Function
- 5 Experimental Evaluation
- 5.1 Experimental Setting
- 5.2 Performance Analysis
- 6 Conclusion
- References
- Uncertainty in GNN Learning Evaluations: The Importance of a Consistent Benchmark for Community Detection
- 1 Introduction
- 2 Related Work
- 3 Methodology
- 3.1 Problem Definition
- 3.2 Hyperparameter Optimisation Procedure
- 3.3 Suite of Tests
- 3.4 Models
- 4 Evaluation and Discussion
- 5 Conclusion
- References
- Link Analysis and Ranking
- Stochastic Degree Sequence Model with Edge Constraints (SDSM-EC) for Backbone Extraction
- 1 Introduction
- 1.1 Preliminaries
- 2 Backbone Models
- 2.1 The Stochastic Degree Sequence Model (SDSM)
- 2.2 The Stochastic Degree Sequence Model with Edge Constraints (SDSM-EC)
- 3 Results
- 3.1 Estimating Q
- 3.2 Toy Illustration
- 3.3 Empirical Illustration
- 4 Conclusion
- References
- Minority Representation and Relative Ranking in Sampling Attributed Networks
- 1 Introduction
- 2 Network Model with Attributes and Homophily
- 3 Network Sampling and Minority Representation
- 3.1 Sampling Methods
- 3.2 Asymptotic Analysis: Sampling in Tree Networks
- 3.3 Synthetic Networks
- 3.4 Real Network
- 4 Relative Ranking of Minorities Under Sampling
- 4.1 Synthetic Networks
- 4.2 Real Networks
- 5 Conclusions and Future Work
- References
- A Framework for Empirically Evaluating Pretrained Link Prediction Models
- 1 Introduction
- 2 Methodology
- 2.1 Features
- 2.2 Training and Testing Set Generation
- 2.3 Stacked Classifier
- 2.4 Cross-Validation Framework
- 3 Experiments
- 3.1 Experimental Setup
- 3.2 Feasibility of Transfer Learning in Link Prediction
- 3.3 Topological Feature Importance for Pretraining Models
- 3.4 Influence of Network Dissimilarity on Transfer Learning
- 4 Conclusion
- References
- Masking Language Model Mechanism with Event-Driven Knowledge Graphs for Temporal Relations Extraction from Clinical Narratives
- 1 Introduction
- 2 Related Work
- 3 Methodology
- 3.1 KG Construction: Event Graph
- 3.2 Text and Graph Embeddings
- 3.3 Language Modelling with Masks (LMM) and Prediction
- 4 Results and Discussion
- 5 Conclusion
- References
- Machine Learning and Networks
- Efficient Approach for Patient Monitoring: ML-Enabled Framework with Smart Connected Systems
- 1 Introduction
- 2 Literature Study and Related Work
- 3 Design and Methodology
- 4 Hardware Design of the System
- 5 Software and Mobile Application Design
- 5.1 Mobile Application
- 5.2 Software and Backend Host
- 6 Data Analysis Framework and Machine Learning Prediction
- 6.1 Linear Regression (LR)
- 6.2 Decision Tree Classifier (DT)
- 6.3 K-Nearest Neighbour Prediction (KNN)
- 7 Work Flow
- 8 System Deployment and Testing
- 9 Conclusions and Future Research
- References
- Economic and Health Burdens of HIV and COVID-19: Insights from a Survey of Underserved Communities in Semi-Urban and Rural Illinois
- 1 Introduction
- 2 Related Work
- 3 Methods
- 3.1 The Burden of HIV Dataset
- 3.2 Data Curation
- 3.3 Graph Inference
- 3.4 Clustering
- 4 Results
- 4.1 Selection Based on African American Race
- 4.2 Selection Based on HIV Positive Status
- 4.3 Selection Based on COVID-19 Influence on Finances
- 5 Discussion and Conclusion
- References
- Untangling Emotional Threads: Hallucination Networks of Large Language Models
- 1 Introduction
- 2 Methodology
- 2.1 Dataset
- 2.2 Models
- 3 Experimental Results
- 3.1 Consistency Analysis of Models
- 3.2 Distance Measures
- 3.3 Clustering Analysis
- 3.4 Confidence Level in Labeling
- 3.5 MultiGraph of Emotions Labeled by LLM Models
- 3.6 Communities Within Vocabulary Networks of LLMs
- 3.7 Networks Metrics
- 4 Conclusion
- References
- Analyzing Trendy Twitter Hashtags in the 2022 French Election
- 1 Introduction
- 2 Related Works
- 3 Data
- 4 Methods
- 4.1 Semantic Network Generation
- 4.2 Semantic User Enrichment
- 4.3 Baseline User Enrichment
- 4.4 Regression Experiment
- 5 Results
- 5.1 Experiment Results
- 5.2 Semantic Analysis
- 6 Conclusions and Future Work
- References
- Rewiring Networks for Graph Neural Network Training Using Discrete Geometry
- 1 Introduction
- 2 Background and Preliminaries
- 2.1 Discrete Geometry and Curvature
- 3 Methods and Experimental Design
- 3.1 Stochastic Discrete Ricci Flow
- 3.2 Datasets
- 3.3 Experimental Setup
- 4 Results: GNN Training with Graph Rewiring
- 4.1 Accuracy
- 4.2 Computational Runtime
- 5 Discussion
- References
- Rigid Clusters, Flexible Networks
- 1 Introduction
- 2 Literature Review
- 3 Methods
- 3.1 Preliminary Definition of Demonstrating a Certainty Effect (DCE)
- 4 Algorithm
- 4.1 Experimentation
- 4.2 Analysis
- 5 Results
- 5.1 Part I - R&F Algorithm Implementation
- 5.2 Part II - Model Application
- 6 Summary and Discussions
- References
- Beyond Following: Augmenting Bot Detection with the Integration of Behavioral Patterns
- 1 Introduction
- 2 Methodology
- 2.1 Dataset
- 2.2 BotRGCN
- 2.3 Derived Relations
- 3 Experiments
- 4 Conclusion
- References
- Graph Completion Through Local Pattern Generalization
- 1 Introduction
- 2 Results
- 2.1 The Network Completion Problem
- 2.2 Experiments
- 2.3 Metrics
- 2.4 Discussion
- References
- A Consistent Diffusion-Based Algorithm for Semi-Supervised Graph Learning
- 1 Introduction
- 2 Dirichlet Problem on Graphs
- 2.1 Heat Equation
- 2.2 Solution to the Dirichlet Problem
- 3 Node Classification Algorithm
- 3.1 Binary Classification
- 3.2 Multi-class Classification
- 4 Analysis
- 4.1 Block Model
- 4.2 Dirichlet Problem
- 4.3 Classification
- 5 Experiments
- 5.1 Synthetic Data
- 5.2 Real Data
- 6 Conclusion
- A Proof of Lemma 1
- B Proof of Theorem 1
- References
- Leveraging the Power of Signatures for the Construction of Topological Complexes for the Analysis of Multivariate Complex Dynamics
- 1 Introduction
- 2 Background on Signatures and Topology
- 2.1 Recalls on the Theory of Signatures
- 2.2 Recalls on Topology
- 3 Building Simplicial Complexes Between Time Series
- 3.1 Presentation of the Method
- 3.2 Our Greedy Order Stratified Algorithm:
- 4 Numerical Experiments and Applications
- 4.1 Practical Choices and Hyper Parameters
- 4.2 Modelling Interactions in Functional MRI Datasets
- 5 Discussion and Future Work
- References
- Empirical Study of Graph Spectra and Their Limitations
- 1 Introduction
- 2 Previous Work
- 3 Mathematical Background
- 3.1 Matrices Under Consideration
- 3.2 Eigenvalues of Normalized Symmetric Laplacian Matrices
- 3.3 Eigengap Heuristic
- 3.4 Graphs Under Consideration
- 4 Empirical Tests
- 4.1 Sensitivity to Graph Size
- 4.2 Sensitivity to Block Size
- 4.3 Sensitivity to Inter-block Edge Probability
- 5 Conclusion and Future Work
- References
- FakEDAMR: Fake News Detection Using Abstract Meaning Representation Network
- 1 Introduction
- 2 Related Work
- 3 Methodology
- 3.1 FakEDAMR: Fake nEws Detection Using Abstract Meaning Representation
- 4 Experimental Setup
- 5 Results
- 6 Ablation Study: Effect of AMR Graph Features
- 7 Conclusion
- References
- Visual Mesh Quality Assessment Using Weighted Network Representation
- 1 Introduction
- 2 Proposed Method
- 2.1 3D Mesh Saliency
- 2.2 Weighted Graph Representation of 3D Mesh
- 2.3 Graph Local Characteristics
- 2.4 Statistical Parameters Estimation
- 2.5 Feature Learning and Regression
- 3 Experimental Results
- 3.1 Datasets
- 3.2 Evaluation Criteria
- 3.3 Effect of Weighting Graph
- 3.4 Comparison with the State-of-the-art
- 4 Conclusion
- References
- Multi-class Classification Performance Improvements Through High Sparsity Strategies
- 1 Introduction
- 2 Materials and Methods
- 2.1 Background and Related Works
- 2.2 Dataset
- 2.3 Methodology
- 3 Discussion and Conclusions
- References
- Learned Approximate Distance Labels for Graphs
- 1 Introduction
- 2 Theoretical Analysis
- 2.1 Problem Definition
- 2.2 Combinatorial Distance Labeling
- 2.3 Approximate Distance Labeling
- 3 Approach
- 3.1 Training Data
- 3.2 Training
- 3.3 Label Quantization Strategies
- 4 Experiments
- 4.1 Dataset
- 4.2 Evaluation Metrics
- 5 Results
- 5.1 Cycles and Alpha Values
- 5.2 Trees
- 5.3 General Graphs
- 6 Limitations
- 7 Computation Specifications
- 8 Relation to Prior Work
- 9 Conclusion
- References
- Investigating Bias in YouTube Recommendations: Emotion, Morality, and Network Dynamics in China-Uyghur Content
- 1 Introduction
- 1.1 Background of Study
- 1.2 The China-Uyghur Crisis
- 2 Methodology
- 2.1 Collection of Data
- 2.2 Emotion Analysis
- 2.3 Morality Assessment
- 2.4 Network Analysis
- 3 Results
- 3.1 Emotion Analysis
- 3.2 Morality Assessment
- 3.3 Network Analysis
- 4 Discussion and Conclusion
- References
- Improving Low-Latency Mono-Channel Speech Enhancement by Compensation Windows in STFT Analysis
- 1 Introduction
- 2 Proposed Method
- 2.1 Spectral Methods: Pipeline and Algorithmic Latency
- 2.2 Compensation Windows in Analysis
- 2.3 Choice of Windowing Function
- 2.4 Multi Encoder Deep Neural Network for Low Latency Deep Noise Suppression
- 3 Experiments
- 3.1 Datasets and Metrics
- 3.2 Training and Model Configurations
- 3.3 Comparison of Different Windowing Strategies Under Different Latency
- 4 Ablation Study
- 5 Conclusions
- References
- Network Embedding
- Filtering Communities in Word Co-Occurrence Networks to Foster the Emergence of Meaning
- 1 Introduction
- 2 SINr: Interpretable Word Vectors Based on Communities
- 3 SINr-filtered: Sampling Communities Using Activations
- 4 Experiments and Results
- 4.1 Experimental Setup
- 4.2 Results
- 5 Conclusion
- References
- DFI-DGCF: A Graph-Based Recommendation Approach for Drug-Food Interactions
- 1 Introduction
- 2 Methodology
- 2.1 Data Preparation
- 2.2 DFI Network Construction
- 2.3 DGCF Model Adaptation
- 2.4 DFIs Prediction
- 3 Evaluation and Results
- 4 Conclusion
- References
- L2G2G: A Scalable Local-to-Global Network Embedding with Graph Autoencoders
- 1 Introduction
- 2 Preliminaries
- 3 Methodology
- 4 Experimental Evaluation
- 5 Conclusion and Future Work
- References
- A Comparative Study of Knowledge Graph-to-Text Generation Architectures in the Context of Conversational Agents
- 1 Introduction
- 2 Background
- 2.1 Graph Embeddings
- 2.2 Mainstream Architectures
- 3 Knowledge Graph-to-Text Generation Architectures
- 3.1 Graph Linearization
- 3.2 Graph Neural Networks (GNNs)
- 3.3 Graphs Transformers (GTs)
- 4 Experiments on Some Models and Datasets
- 4.1 Datasets and Metrics
- 4.2 Experiments
- 5 Results and Discussion
- 6 Conclusion and Perspectives
- References
- Network Embedding Based on DepDist Contraction
- 1 Introduction
- 2 Related Work
- 3 Non-symmetric Structural Dependency
- 3.1 Distance Based on Non-symmetric Dependency
- 4 DepDist Contraction
- 4.1 Algorithm
- 4.2 One Step of Iteration
- 4.3 Selecting Node to Move
- 4.4 When to Stop Iterating
- 4.5 Scalability
- 5 Experiment
- 5.1 Results
- 6 Conclusion and Future Work
- References
- Evaluating Network Embeddings Through the Lens of Community Structure
- 1 Introduction
- 2 Background
- 2.1 Random-Walk-Based Methods
- 2.2 Matrix Factorization-Based Methods
- 2.3 Deep Learning-Based Method
- 2.4 Evaluation Metrics
- 3 Experimental Setup and Evaluation
- 3.1 Synthetic Network Generation
- 3.2 Experimental Setup
- 3.3 Voting Model
- 4 Results and Discussion
- 5 Conclusion
- References
- Deep Distance Sensitivity Oracles
- 1 Introduction
- 1.1 Contributions
- 1.2 Related Work
- 2 Theoretical Analysis
- 3 Method
- 3.1 Training Data
- 3.2 Representation Learning with Graph Convolutional Networks
- 3.3 Multi-layer Perceptron
- 3.4 Summary
- 4 Experiments
- 4.1 Datasets
- 4.2 Evaluation
- 5 Results
- 6 Conclusion
- References
- Correction to: Leveraging the Power of Signatures for the Construction of Topological Complexes for the Analysis of Multivariate Complex Dynamics
- Correction to: Chapter 24 in: H. Cherifi et al. (Eds.): Complex Networks & Their Applications XII, SCI 1141, https://doi.org/10.1007/978-3-031-53468-3_24
- 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.