Complex Networks & Their Applications XII

Name: Complex Networks & Their Applications XII | Proceedings of The Twelfth International Conference on Complex Networks and their Applications: COMPLEX NETWORKS 2023 Volume 1
Brand: Springer
Price: 223.63 EUR
Availability: OnlineOnly

Proceedings of The Twelfth International Conference on Complex Networks and their Applications: COMPLEX NETWORKS 2023 Volume 1

Hocine Cherifi Luis M. Rocha Chantal Cherifi Murat Donduran(Herausgeber*in)

Springer (Verlag)

Erschienen am 19. Februar 2024

XXII, 466 Seiten

E-Book

PDF mit Wasserzeichen-DRM

Systemvoraussetzungen

978-3-031-53468-3 (ISBN)

223,63 €inkl. 7% MwSt.

Systemvoraussetzungen

für PDF mit Wasserzeichen-DRM

E-Book Einzellizenz

Als Download verfügbar

Beschreibung

Weitere Details

Weitere Ausgaben

Personen

Luis M. Rocha is the George J. Klir Professor of Systems Science at the Thomas J. Watson College of Engineering and Applied Science (Department of Systems Science and Industrial Engineering), Binghamton University (State University of New York), where he leads the Complex Adaptive Systems and Computational Intelligence (CASCI) lab. He is also Senior Fellow at the Instituto Gulbenkian da Ciencia, co-director of the Consortium for Social and Biomedical Complexity between Binghamton University and Indiana University, and a founding partner of the international Center for Excellence in Mental Health Sciences. He has been the director of the NSF-NRT Interdisciplinary Training Program in Complex Networks and Systems, and Professor of Informatics in the Luddy School of Informatics, Computing, and Engineering at Indiana University, where he was a member of the advisory council of the Indiana University Network Science Institute, and core faculty of the Cognitive Science Program. Dr. Rocha is a Fulbright Scholar and has been a Visiting Professor a the Center for Theoretical Physics at the Aix-Marseille University, France as well as the Neuroscience Program, at the Champalimaud Foundation and the NOVA School of Business and Economics in Portugal. His research is on complex networks & systems, computational & systems biology, and computational intelligence. He received his Ph.D in Systems Science in 1997 from the State University of New York at Binghamton. From 1998 to 2004, he was a permanent staff scientist at the Los Alamos National Laboratory, where he founded and led a Complex Systems Modeling Team from 1998 to 2002 and was part of the Santa Fe Institute research community. He has organized major conferences in the field, such as the Tenth International Conferenceon the Simulation and Synthesis of Living Systems (Alife X), the Ninth European Conference on Artificial Life (ECAL 2007), and Complex Networks 2019-2023. He has published many articles in scientific and technology journals and has been the recipient of several scholarships and awards. At Indiana University, he has received the Indiana University School of Informatics & Computing, Trustees Award for Teaching Excellence in 2006 and 2015 after developing the complex systems training program and syllabi for several courses.

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.

Inhalt

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

Systemvoraussetzungen

Als PDF speichern Als Link merken