دانلود کتاب Artificial Intelligence Applications and Innovations: 18th IFIP WG 12.5 International Conference, AIAI 2022 Hersonissos, Crete, Greece, June 17–20, 2022 Proceedings, Part II

فهرست مطالب :

Preface Organization Contents – Part II Contents – Part I Fuzzy Modeling and IoT A Communication Data Layer for Distributed Neuromorphic Systems 1 Introduction 2 Neuromorphic Data Layer 2.1 Raw Spike API 2.2 Symbolic Representation API 2.3 Encoding of Integer Valued State Variables 2.4 Practical Integer Encoding and Decoding on Neuromorphic Hardware 3 Distributed Neuromorphic Prototype System 3.1 End-to-End Use-Case 3.2 Heterogeneity Aspects of the Prototype System 4 Visual Analysis and Data Layer on Loihi 5 Empirical Results 6 Conclusion References Brainstorming Fuzzy Cognitive Maps for Camera-Based Assistive Navigation 1 Introduction 2 Theoretical Background 2.1 Fuzzy Cognitive Maps 2.2 Determinative Brain Storm Optimization 3 Brainstorming Cognitive Maps 4 Experiments and Simulation Results 5 Conclusion References Creating a Bridge Between Probabilities and Fuzzy Sets and Its Impact on Drought Severity Assessment 1 Introduction 2 Fuzzy Methodology 2.1 Fuzzy Sets 2.2 Fuzzy Estimators 2.3 Proposed Methodology 3 Application and Discussion 4 Concluding Remarks References SAF: A Peer to Peer IoT LoRa System for Smart Supply Chain in Agriculture 1 Introduction 2 Related Work 3 SAF Description 3.1 Objectives and Design 3.2 Wireless Network Protocol Comparisons 4 Analytics and Events 5 Conclusions References Machine Learning Classification A Multi-label Time Series Classification Approach for Non-intrusive Water End-Use Monitoring 1 Introduction 2 Problem Formulation 3 Experimental Setup 3.1 Simulator and Dataset 3.2 Classification Algorithms 3.3 Multi-task Classification Methods 3.4 Evaluation Metrics 4 Empirical Analysis 4.1 Hyper-Parameter Tuning 4.2 Role of the Sliding Window Size 4.3 Role of the Multi-label Classification Method 5 Comparative Study 6 Conclusions and Future Work References A Primer for tinyML Predictive Maintenance: Input and Model Optimisation 1 Introduction 2 Related Work 3 Background of Applying tinyML 4 Model Optimisations 5 Input Optimisations 6 Conclusion References Allocating Orders to Printing Machines for Defect Minimization: A Comparative Machine Learning Approach 1 Introduction 2 Methodology and Algorithmic Approach 2.1 Available Dataset 2.2 Machine Selection Framework 2.3 ML Algorithms for Accuracy Prediction 3 Numerical Results 3.1 Training of the Multi-layer Perceptron 3.2 Ensemble Learning Techniques 3.3 Comparison Between ML Methods 3.4 Evaluation of the Proposed Machine Selection Policy 4 Conclusion References Bias in Face Image Classification Machine Learning Models: The Impact of Annotator's Gender and Race 1 Introduction 2 Background and Literature Review 2.1 Face Image Interpretation 2.2 Bias in Computer Vision Algorithms 3 Face Database and Annotation 4 Experiment 1: Comparing the Performance of Annotator-Specific Classification Models 4.1 Model Training 4.2 Results and Discussion 5 Experiment 2: Predicting Annotator Groups Based on Annotations 6 Conclusions and Future Work References Decision Tree Induction Through Meta-learning 1 Introduction 2 A Meta Decision Tree Algorithm 2.1 Leaf-Weighting Metrics 2.2 Tree Leaves Weights Normalization 2.3 Example 3 Experimental Setup 4 Results and Discussion 5 Conclusions References Hybrid (CPU/GPU) Exact Nearest Neighbors Search in High-Dimensional Spaces 1 Introduction 2 Related Work 3 Algorithm 3.1 Distance Computation 3.2 k-Selection 3.3 Batching 4 Results 5 Conclusions References Machine Learning Approach to Detect Malicious Mobile Apps 1 Introduction 2 Features Used in Simulation 3 Results and Evaluations 4 Conclusion References Prediction of Wafer Map Categories Using Wafer Acceptance Test Parameters in Semiconductor Manufacturing 1 Introduction 2 Methodology 2.1 Data Preprocessing 2.2 Oversampling Algorithm 2.3 Machine Learning Algorithm 3 Experimental Studies 3.1 Experiment Setup 3.2 Model Training 3.3 Experiment Results 4 Conclusions References Machine Learning Modeling /Feature Selection An Improved Neural Network Model for Treatment Effect Estimation 1 Introduction 2 Related Work 3 Modified Dragonnet Model 3.1 Calculation of Average Outcome Vectors 3.2 Modified Dragonnet Architecture 4 Data 5 Experimental Results 6 Conclusion References An Industry 4.0 Intelligent Decision Support System for Analytical Laboratories 1 Introduction 2 Related Work 3 Materials and Methods 3.1 Problem Formulation 3.2 Proposed IDSS 3.3 Evaluation 4 Results 4.1 Developed IDSS Prototype 4.2 Evaluation 5 Conclusions References Combining Cox Model and Tree-Based Algorithms to Boost Performance and Preserve Interpretability for Health Outcomes 1 Introduction 2 Methods 2.1 Models 2.2 Samples 2.3 Performance Assessment 3 Results 4 Discussion References Distribution Guided Neural Disaggregation of PM10and O3Hourly Concentrations from Daily Statistics and Low-Cost Sensors 1 Introduction 1.1 Air Quality Monitoring 1.2 LCS Calibration as Time Series Nowcasting 1.3 Aim of the Study 2 Materials and Methods 2.1 Experimental Set-Up 2.2 Preprocessing and Temporal Feature Engineering 2.3 Distribution Guided Neural Disaggregation 2.4 Evaluation 3 Results 4 Conclusions References Experimental Comparison of Metaheuristics for Feature Selection in Machine Learning in the Medical Context 1 Introduction 2 Methodology Used 2.1 Evaluation Criteria 2.2 Obtaining and Validating an `Optimal' Subset 3 Feature Selection with the Use of Metaheuristics 3.1 Solution-Based Metaheuristics 3.2 Population-Based Metaheuristics 3.3 A New Population-Based Metaheuristics 4 Datasets 4.1 ALS Database 4.2 Benchmarks 5 Results 6 Conclusion References Exploring the Pertinence of Distance Functions for Nominal Multi-label Data 1 Introduction 2 Related Works 3 A Toy Example and the Motivation of This Work 4 Exploration 4.1 The Proposed Ensemble 5 Experimental Study 6 Results and Analysis 7 Conclusion References Feature Selection Methods for Uplift Modeling and Heterogeneous Treatment Effect 1 Introduction 2 Related Work 3 Uplift Models 4 Feature Selection Methods for Uplift Modeling 4.1 Filter Methods 4.2 Embedded Methods 5 Empirical Evaluation 5.1 Experiment 1: Evaluation with Synthetic Data 5.2 Experiment 2: Evaluation with MegaFon Uplift Competition Data 6 Discussion 7 Conclusion References Machine Learning Applications in Real Estate: Critical Review of Recent Development 1 Introduction 2 Materials and Methodology 3 ML/DL Algorithms and Their Applications in Rea Estate 3.1 Neural Networks (NN) 3.2 Random Forest, Bagging and Decision Tree 3.3 Boosting and Gradient Boosting 3.4 M5 Model Tree and Cubist 3.5 Support Vector Machine (SVM) 3.6 Linear and Non-linear Regression Model 4 Results and Discussion 5 Conclusions and Future Prospects References Predictive Maintenance Based on Machine Learning Model 1 Introduction 2 Related Works 3 Problem Statement 4 Overall Proposed Architecture 4.1 Machine Learning Model 4.2 Architecture 4.3 Technical Set up 5 Results and Discussion 6 Conclusion References Production Time Prediction for Contract Manufacturing Industries Using Automated Machine Learning 1 Introduction 2 Background 3 Materials and Methods 3.1 CRISP-DM 3.2 AutoML 4 CRISP-DM Methodology 4.1 First CRISP-DM Iteration 4.2 Second CRISP-DM Iteration 4.3 Third CRISP-DM Iteration 5 Discussion 6 Conclusion References Social Media, Sentiment Analysis/Natural Language - Text Mining A Multi-Objective Optimization Algorithm for Out-of-Home Advertising 1 Introduction 2 Data and Feature Engineering 2.1 GIS Data 2.2 Demographics Dataset 2.3 Billboards Dataset 2.4 Feature Engineering 3 Problem Formulation 4 Multi-Objective Evolutionary Algorithm Based on Decomposition (MOEA/D) 4.1 Problem-Specific Heuristics 5 Experimental Setup and Results 5.1 Experimental Setup 5.2 Experimental Study 1: MOO for OOH Advertising 5.3 Experimental Study 2: MOEA/D Versus MOEA/D-i 5.4 Experimental Study 3: MOEA/D Versus MOEA/D-r 6 Conclusions References AutoMC: Learning Regular Expressions for Automated Management Change Event Extraction from News Articles 1 Introduction 2 Related Work 3 AutoMC 3.1 Regex Learner 3.2 News Classifier 3.3 Management Change Extractor 4 Evaluations and Experiments 5 Conclusion References How Dimensionality Reduction Affects Sentiment Analysis NLP Tasks: An Experimental Study 1 Introduction 2 Related Work 3 Sentiment Analysis and Dimensionality Reduction 3.1 Datasets 3.2 Text Classifiers 3.3 Text Preprocessing and Dimensionality Reduction 3.4 Results 4 Discussion 5 Conclusions and Future Work References Invention Concept Latent Spaces for Analogical Ideation 1 Introduction 1.1 Analogy in Latent Spaces 1.2 Creativity Using Latent Spaces 2 Text Autoencoders 2.1 The Denoising Autoencoder (DAE) 2.2 The Variational Autoencoder (VAE) 2.3 β-VAE 3 Problems with Latent Vectors in Text VAEs 3.1 Latent Space Collapse 3.2 ‘Prompt Only’ Text Generation from Latent Vector 4 Hierarchical VAE for Text 5 Syntax and Semantics 6 Implementation 6.1 Data 6.2 Network Architecture 6.3 Training 7 Results 7.1 Concept Fusion 7.2 Latent White Space 7.3 Citation Analogy 7.4 Evaluation 8 Summary References Multilingual Sentiment Analysis on Twitter Data Towards Enhanced Policy Making 1 Introduction 2 Related Work 3 Multilingual Sentiment Analysis 3.1 Multilingual Sentence Classification (MSC) 3.2 Zero-Shot Sentiment Classification (ZSSC) 4 Experimental Results 4.1 Datasets Description 4.2 Evaluation Setup and Measures 4.3 Pre-processing 4.4 Results and Discussion 5 Conclusion References On the Evaluation of the Plausibility and Faithfulness of Sentiment Analysis Explanations 1 Introduction 2 Background 2.1 Sentiment Analysis Models 2.2 Explainability Models 3 Evaluation Dataset 4 Methodology 4.1 Experimental Design 4.2 Plausibility Evaluation Metrics 5 Results and Discussion 5.1 Back-end Setup 5.2 Evaluation of Faithfulness 5.3 Plausibility Evaluation on Ground Truth Data 5.4 Qualitative Evaluation 5.5 Take-Away Messages 6 Conclusion References Sentiment Analysis on COVID-19 Twitter Data: A Sentiment Timeline 1 Introduction 2 Related Work 3 Methodology 3.1 Dataset 3.2 Sentiment Model Development 3.3 Sentiment Analysis 4 Results 4.1 Overview 4.2 Twitter Timeline 4.3 Sentiment Overview 4.4 Sentiment Timeline 4.5 Influential Power 5 Conclusion 5.1 Discussion 5.2 Future Work References Social Media Sentiment Analysis Related to COVID-19 Vaccines: Case Studies in English and Greek Language 1 Introduction 2 Background 2.1 Sentiment Analysis 2.2 Sentiment Analysis for Multilingual Documents 3 Research Design 3.1 Dataset Collection and Annotation 3.2 Data Pre-processing 3.3 Text Representation and Classification 4 Results 4.1 Trend Analysis 5 Conclusion 5.1 Limitations and Future Work References Time Series Modeling/Transfer Learning Comparing Boosting and Deep Learning Methods on Multivariate Time Series for Retail Demand Forecasting 1 Introduction 2 Background 3 Model Analysis 3.1 Multivariate Boosting 3.2 TCN 3.3 TFT 4 Performance Evaluations 4.1 The Dataset 4.2 Model Fine-Tuning and Training 4.3 Results 5 Conclusions References Equilibrium Resolution for Epoch Partitioning 1 Introduction 2 Literature Review 3 Preliminaries 3.1 Prolonged Events 3.2 Epoch Partitioning 4 Equilibrium Resolution: A Numerical Example 5 Probabilistic Modelling 5.1 The Probabilistic Approach to Equilibrium Resolution 5.2 Approximating Equilibrium Resolution 5.3 Equilibrium Resolution for Lognormaly Distributed Agents Action Duration 5.4 Equilibrium Resolution for Truncated Lognormal Distributed Agents Action Duration 6 Application on Case Study Data 6.1 Case Study: Trading Data 6.2 Case Study: NYC Taxi Journey Data 7 Partitioning Based on Monitoring Functions 8 Conclusion References Topological Data Analysis of Time-Series as an Input Embedding for Deep Learning Models 1 Introduction 2 Background 2.1 Time-Series Embeddings 2.2 Takens' Embedding 2.3 Topological Data Analysis (TDA) 2.4 Persistence Diagrams 3 Methodology 3.1 Datasets 3.2 Tools and Libraries 3.3 Data Pipeline 4 Results 5 Conclusion References Transfer Learning for Predicting Gene Regulatory Effects of Chemicals 1 Introduction 2 Materials and Methods 2.1 Datasets 2.2 Feature Generation 2.3 Data Splitting 2.4 Experiments 3 Discussion 3.1 Best Improvements 3.2 Complications of Using Random Split 3.3 Random-Split in Source Model Training 3.4 Splitting Up and Down Regulations 4 Conclusion References Transfer Learning with Jukebox for Music Source Separation 1 Introduction and Related Work 2 Method 2.1 Architecture 2.2 Data 2.3 Training 3 Experiments and Results 4 Conclusion References Unsupervised Modeling An Inductive System Monitoring Approach for GNSS Activation 1 Introduction 2 Related Work 3 Methodology 3.1 An Inductive System Monitoring Approach 3.2 The Proposed ISM-Based Approach for GNSS Activation 4 Experimental Design 4.1 Dataset 4.2 Data Preprocessing 4.3 Evaluation Metrics 5 Results and Discussion 6 Conclusions and Future Work References Client Segmentation of Mobile Payment Parking Data Using Machine Learning 1 Introduction 2 Data Collection and Preprocessing 2.1 Attribute Extraction 2.2 Data Preprocessing 3 Experimental Results 3.1 K-Means Clustering 3.2 DBScan Clustering 4 Conclusions References Determining Column Numbers in Résumés with Clustering 1 Introduction 2 Data Set 3 Methodology 3.1 K-means Algorithm 3.2 Elbow Method 3.3 Silhouette Method 3.4 DBSCAN Algorithm 3.5 Column Detection via Clustering 4 Experiments and Results 5 Conclusion References High Rank Self-Organising Maps for Image Fingerprinting 1 Introduction 2 Self Organising Maps 3 Method 3.1 Fingerprint Generation 3.2 Fingerprint Comparison 4 Results 4.1 MNIST 4.2 Noise Resilience 5 Conclusion References Implicit Maximum Likelihood Clustering 1 Introduction 2 Generative Modeling Using IMLE 3 Clustering Based on IMLE 3.1 Cluster Friendly Input Distribution 3.2 Two-Stage Nearest Neighbor Search for Determining Data Representatives 4 Experiments 4.1 Datasets 4.2 Neural Network Architecture 4.3 Evaluation Metrics 4.4 Experimental Results 5 Conclusions References Query Driven Data Subspace Mapping 1 Introduction 2 Related Work 3 Preliminaries 4 The Proposed Approach 5 Experimental Setup and Evaluation 6 Conclusions and Future Work References Correction to: Transfer Learning with Jukebox for Music Source Separation Correction to: Chapter “Transfer Learning with Jukebox for Music Source Separation” in: I. Maglogiannis et al. (Eds.): Artificial Intelligence Applications and Innovations, IFIP AICT 647, https://doi.org/10.1007/978-3-031-08337-2_35 Author Index