Explainable Artificial Intelligence for Smart Cities

Cover
Half Title
Title Page
Copyright Page
Contents
Contributors
1. An Overview of Explainable Artificial Intelligence (XAI) from a Modern Perspective
	1.1. Introduction
	1.2. Explainable Artificial Intelligence (XAI) Concept
	1.3. Need for XAI in Neural Network-Oriented Applications
	1.4. Discussion
	1.5. Trends
	1.6. Conclusions
	References
2. Explainable Artificial Intelligence for Services Exchange in Smart Cities
	2.1. Introduction
	2.2. Explainable AI and Decision-Making
	2.3. Big Data and ICT for XAI Services in the Context of Smart Cities
	2.4. The Smart Environment with XAI
		2.4.1. Proposed Smart City Main Pillars and Architecture
			2.4.1.1. Transition Entities
				Government Entities
				Private Sector Investors
				Ad hoc Research Units
			2.4.1.2. Policies Standardization and Protocols
			2.4.1.3. Applications Layer Structure Considering XAI
	2.5. Benefits of Using XAI in Smart Cities
	2.6. Addressed Challenges for XAI Data Creation in Smart Cities
	2.7. Summary
	References
3. IoT- and XAI-Based Smart Medical Waste Management
	3.1. Introduction
	3.2. Related Work
		3.2.1. Medical Waste Management as an Application of Internet of Things-Based Smart Cities
		3.2.2. Explainable Artificial Intelligence for Smart Cities
	3.3. Proposed Approach
		3.3.1. The Current Medical Waste Disposal Process in Morocco
			3.3.1.1. Legislative Framework for Medical Waste Management in Morocco
			3.3.1.2. The Current Model of Medical Waste Disposal
		3.3.2. The Proposed Smart Model for Medical Waste Management
		3.3.3. Architecture of the Proposed Smart Solution
			3.3.3.1. Overview
			3.3.3.2. Smart Devices
			3.3.3.3. IoT Gateway
			3.3.3.4. Data Preprocessing Layer
			3.3.3.5. Data Processing Layer
			3.3.3.6. Application Layer
	3.4. Expected Results
	3.5. Conclusion
	References
4. The Impact and Usage of Smartphone among Generation Z: A Study Based on Data Mining Techniques
	4.1. Introduction
	4.2. Literature Review
	4.3. Methodology
		4.3.1. Overview
		4.3.2. KDD Process
		4.3.3. Data Mining Methods
			4.3.3.1. Association Rule
			4.3.3.2. Classification
			4.3.3.3. Weka Data Mining Tool
			4.3.3.4. Approach
		4.3.4. Implementation
			4.3.4.1. Data Cleaning
			4.3.4.2. Data Transformation
			4.3.4.3. Generalization
			4.3.4.4. Data Reduction
			4.3.4.5. Attribute Subset Selection
			4.3.4.6. Discretization
			4.3.4.7. Association Rules
	4.4. Results and Discussion
		4.4.1. Data Collection
		4.4.2. Data Summarization
			4.4.2.1. Distribution of Gender
			4.4.2.2. Distribution of Age
			4.4.2.3. Distribution of Marital Status
			4.4.2.4. Distribution of Education Level (Higher Education)
			4.4.2.5. Distribution of Field Which Related to Information Technology or Non-Information Technology
			4.4.2.6. Distribution of Respondents' Awareness Regarding the Side Effects of Smartphone Usage
			4.4.2.7. Respondents' Opinion of Smartphone Use by Their Kids
			4.4.2.8. Time Period Taken for the above Decision
			4.4.2.9. Respondents' Opinion about Smartphone Technology Necessary to Children's Lives or Not
			4.4.2.10. Distribution of Respondents That Worry about the Society Underestimating When Their Child Doesn't Know How to Handle a Smartphone
			4.4.2.11. Respondents' Opinion on Monitoring Children's Mobile Activities
			4.4.2.12. Type of Mobile Activities and Monitoring Rate
			4.4.2.13. Respondents' Distribution in Preventing or Limiting Children's Smartphone Usage
			4.4.2.14. Respondents' Opinion on Children Socializing More by Using Smartphones via Social Media
		4.4.3. Evaluation
		4.4.4. Classification
	4.5. Conclusion
	References
5. Explainable Artificial Intelligence: Guardian for Cancer Care
	5.1. Introduction
	5.2. Explainable Artificial Intelligence
		5.2.1. Evolution of Explainable Artificial Intelligence
		5.2.2. Artificial Intelligence: Mechanism
	5.3. Artificial Intelligence in Cancer Treatment and Drug Discovery
		5.3.1. Drug Discovery
		5.3.2. Prediction of Oncogenesis
		5.3.3. Drug Repurposing
		5.3.4. Modulation and Set Up Therapy Plans
	5.4. Future Prospects
	5.5. Conclusion
	References
6. ANN-Based Brain Tumor Classification: Performance Analysis Using K-Means and FCM Clustering With Various Training Functions
	6.1. Introduction
	6.2. Proposed Methodology
		6.2.1. Preprocessing
			6.2.1.1. Resizing
			6.2.1.2. Sharpening Filter
		6.2.2. Image Segmentation
			6.2.2.1. K-Means Clustering
			6.2.2.2. Fuzzy C-Means Clustering
		6.2.3. Feature Extraction and Reduction
		6.2.4. Classification
			6.2.4.1. Levenberg-Marquardt
			6.2.4.2. BFGS Quasi-Newton
			6.2.4.3. Resilient Backpropagation
			6.2.4.4. Scaled Conjugate Gradient
			6.2.4.5. Conjugate Gradient With Powell-Beale Restarts
			6.2.4.6. Fletcher-Powell Conjugate Gradient
			6.2.4.7. Polak-Ribiére Conjugate Gradient
			6.2.4.8. One Step Secant
			6.2.4.9. Variable Learning Rate Backpropagation
		6.2.5. Algorithm of Two Proposed Approaches
			6.2.5.1. Algorithm 1
			6.2.5.2. Algorithm 2
	6.3. Experimental Results and Analysis
		6.3.1. Dataset
		6.3.2. Classification Accuracy of the Two Proposed Methods
		6.3.3. Analysis of the Best-Obtained Accuracy
		6.3.4. Result Analysis of Algorithm 1 and Algorithm 2
		6.3.5. Comparison of Proposed Work with Relevant Works
	6.4. Conclusion
	Acknowledgements
	References
7. Recognition of the Most Common Trisomies through Automated Identification of Abnormal Metaphase Chromosome Cells
	7.1. Introduction
		7.1.1. Background
		7.1.2. Contributions
		7.1.3. Chapter Organization
	7.2. Related Works
	7.3. Methods
		7.3.1. Image Preprocessing and Segmentation
		7.3.2. Feature Extraction
			7.3.2.1. Chromosome Length
			7.3.2.2. Relative Length
			7.3.2.3. Chromosome Area
			7.3.2.4. Relative Area
			7.3.2.5. Centromere Index (CI)
			7.3.2.6. Density Profile
		7.3.3. Chromosome Classification
		7.3.4. Trisomy Detection
	7.4. Experimental Evaluation and Results
		7.4.1. Dataset
		7.4.2. Performance Measurement
		7.4.3. Experiments and Results
			7.4.3.1. Experiment 1: Test (Classic and Multiple Autoencoder) Neural Network to Classify Chromosomes into Six Class
			7.4.3.2. Experiment 2: Test Multiple Autoencoder Neural Network to Classify Chromosomes into Five Classes
	7.5. Conclusion and Future Work
	References
8. Smart Learning Solutions for Combating COVID-19
	8.1. Introduction
		8.1.1. Background: Smart Education an Essential Component of Smart Cities
		8.1.2. Teaching-Learning Process
		8.1.3. Teaching Pedagogies
			8.1.3.1. Lectures
			8.1.3.2. Worked Examples
			8.1.3.3. Interactive Learning
			8.1.3.4. Spaced-Learning
			8.1.3.5. Flipped Classrooms
			8.1.3.6. Socratic Questioning
			8.1.3.7. Discussion-Based Learning
			8.1.3.8. Case-Based Learning
			8.1.3.9. Collaborative Learning
			8.1.3.10. Enquiry-Based Learning
			8.1.3.11. Problem-Based Learning
			8.1.3.12. Project-Based Learning
			8.1.3.13. Self-Learning
			8.1.3.14. Game-Based Learning and Gamification
			8.1.3.15. VAK Teaching
			8.1.3.16. Cross-Over Learning
	8.2. From Face-to-Face Teaching to Online Teaching
		8.2.1. Computer-Aided Teaching-Learning
		8.2.2. Digital Education
			8.2.2.1. Time Line of Digital Education
			8.2.2.2. Advantages of Digital Learning
	8.3. Initiatives Taken by Governments of Various Countries
		8.3.1. Afghanistan
		8.3.2. Argentina
		8.3.3. Austria
		8.3.4. Bangladesh
		8.3.5. Belize
		8.3.6. Bermuda
		8.3.7. China
		8.3.8. United States
		8.3.9. United Kingdom
		8.3.10. India
	8.4. Smart Technologies for Online Learning
		8.4.1. Mobile Learning
		8.4.2. Microlearning
		8.4.3. Internet of Things
		8.4.4. Cloud-Based e-Learning
		8.4.5. Gamification
		8.4.6. Adaptive e-Learning
		8.4.7. Augmented Reality
		8.4.8. Video e-Learning
		8.4.9. Beacon e-Learning
		8.4.10. Artificial Intelligence e-Learning
	8.5. Smart Resources for Online Teaching, Learning, and Evaluation
		8.5.1. Dropbox
		8.5.2. Class Dojo
		8.5.3. Edmodo
		8.5.4. Educreations
		8.5.5. TED Ed
		8.5.6. Unplag
		8.5.7. Slack
		8.5.8. Google Apps for Education
		8.5.9. Remind
		8.5.10. Edublogs
		8.5.11. Socrative
		8.5.12. Moodle
		8.5.13. Discord
	8.6. Artificial Intelligence-Based Learning and the Emergence of the Intelligent Tutoring System (ITS)
	8.7. Explainable Artificial Intelligence (XAI)
	8.8. Blended Learning Model for Future
		8.8.1. Face-to-Face Driver Model
		8.8.2. Rotation Model
			8.8.2.1. Station Rotation
			8.8.2.2. Lab Rotation
			8.8.2.3. Flipped Rotation
			8.8.2.4. Individual Rotation Model
		8.8.3. Flex Model
		8.8.4. Online Lab School Model
		8.8.5. Self-Blended Model Evaluation of classification methods and learning logs
		8.8.6. Online Driver Model
	8.9. Limitations and Future Prospects
	Notes
	References
9. An Analysis of Machine Learning for Smart Transportation System (STS)
	9.1. Introduction
	9.2. Evolution of Smart Transport System (STS)
	9.3. Process of Smart Transport System
		9.3.1. Why Is a Smart Transportation System Required?
		9.3.2. How Does a Smart Transportation System Work?
			9.3.2.1. STS Joining Innovations
			9.3.2.2. STS Technological Facilitation
			9.3.2.3. Information Acquisition
			9.3.2.4. Information Processing
			9.3.2.5. Information Communications
			9.3.2.6. Information Sharing
			9.3.2.7. Information Exploitation
		9.3.3. Intelligent Transportation System User Functions
	9.4. Need for ML Techniques in STS
	9.5. DL Techniques for Autonomous Vehicle Decision Making
		9.5.1. DL for Driving Scene Perception and Localization
		9.5.2. DL Neural Techniques for Autonomous Driving
		9.5.3. DL for Passage Preparation and Performance Calculation
	9.6. Integration of ML with IoT in Autonomous Vehicles
	9.7. Conclusion
	References
10. Classification of Kinematic Data Using Explainable Artificial Intelligence (XAI) for Smart Motion
	10.1. Introduction
	10.2. Research Review
	10.3. Materials and Methods
		10.3.1. Study Sample
		10.3.2. Experimental Set-Up
		10.3.3. Data Acquisition
		10.3.4. Data Management
		10.3.5. System Evaluation
	10.4. Result
		10.4.1. Whole-Body Movement
		10.4.2. Upper-Limbs Movement
	10.5. Discussion
	10.6. Conclusion
	Acknowledgements
	References
11. Smart Urban Traffic Management for an Efficient Smart City
	11.1. Introduction
	11.2. Smart Transport in Smart City Environment
		11.2.1. Explainable Artificial Intelligence (XAI)
		11.2.2. Edge and Fog Computing for Smart City
		11.2.3. Fog Data as a Service Delivery Model
		11.2.4. Toward a Fog-Based Real-Time Big Data Pipeline
		11.2.5. Smart Transportation Systems (STS) and Vehicular Fog Computing
		11.2.6. Deep Learning (DL) Methods
	11.3. Related Works on Urban Traffic Management UTM
		11.3.1. Urban Traffic Management
		11.3.2. Urban Traffic Management Approaches
		11.3.3. Traffic Lights Management
		11.3.4. DL Approaches for Urban Traffic Management
			11.3.4.1. Traffic Analysis and Prediction
			11.3.4.2. Autonomous Driving
			11.3.4.3. Traffic Signal Control
	11.4. Network for Urban Data Collection
		11.4.1. Internet of Things (IoT)
		11.4.2. Air Pollution Measurements
	11.5. Our Project Architecture
		11.5.1. Hardware Implementation for Data Collection
		11.5.2. Urban Traffic Model
	11.6. Our Approach for Urban Traffic Flow Management
		11.6.1. Concept
		11.6.2. Mathematical Models of Route Flow and Crossroads
			11.6.2.1. Basic Process Model
			11.6.2.2. Fluid Modelling of Road Network Flows
			11.6.2.3. Dynamics of a Fluid Reservoir
		11.6.3. Traffic Lights Management
		11.6.4. Implementation Example
	11.7. Impact of UTM on Urban Supply Chain
		11.7.1. Urban SC towards GrSC in the Literature
		11.7.2. Urban Air Pollution and the GrSCM
		11.7.3. Improvement of Key Indicators of Urban Supply Chain
	11.8. Conclusion
	References
12. Systematic Comparison of Feature Selection Methods for Solar Energy Forecasting
	12.1. Introduction
	12.2. Related Work
	12.3. Feature Selection Algorithms
		12.3.1. Least Absolute Shrinkage and Selection Operator
		12.3.2. Random Forests FS
		12.3.3. Stepwise Regression
	12.4. Real-World Application
		12.4.1. Dataset
		12.4.2. Feature Selection Results and Discussion
	12.5. Conclusion
	References
13. Indoor Environment Assistance Navigation System Using Deep Convolutional Neural Networks
	13.1. Introduction
	13.2. Related Work
	13.3. Recognizing Indoor Objects: Approach Adopted Based on Deep CNN
		13.3.1. Indoor Objects Recognition Using EfficientNet
		13.3.2. Indoor Objects Recognition Using Inception Family
		13.3.3. Indoor Objects Recognition Using ResNet
	13.4. Experiments and Results
	13.5. Conclusion
	References
14. Pixel-Based Classification of Land Use/Land Cover Built-Up and Non-Built-Up Areas Using Google Earth Engine in an Urban Region (Delhi, India)
	14.1. Introduction
		14.1.1. Background
		14.1.2. Methodology
		14.1.3. Chapter Organization
	14.2. Related Work
	14.3. Problem Statement
	14.4. Methodology
		14.4.1. Google Earth Engine (GEE)
		14.4.2. Classification Models
			14.4.2.1. Classification and Regression Tree (CART) or Decision Tree
			14.4.2.2. Support Vector Machine (SVM)
			14.4.2.3. Random Forest
		14.4.3. Cross-Validation
	14.5. Proposed Method
	14.6. Experiment
		14.6.1. Study Area
		14.6.2. Dataset
		14.6.3. Preprocessing and Scene Selection
			14.6.3.1. Top of Atmosphere (TOA) Reflectance
			14.6.3.2. Solar Zenith Angle
			14.6.3.3. Sun Elevation Angle
			14.6.3.4. Normalized Difference Vegetation Index (NDVI)
			14.6.3.5. Normalized Difference Built-Up Index (NDBI)
		14.6.4. Detection of Built-Up Regions
	14.7. Result and Analysis
	14.8. Discussion and Conclusion
	14.9. Future Work Recommendations
	Acknowledgements
	References
15. Emergence of Smart Home Systems Using IoT: Challenges and Limitations
	15.1. Introduction
		15.1.1. Background
		15.1.2. Methodology
		15.1.3. Chapter Organization
	15.2. Functions of Smart Home System
		15.2.1. Alert and Sensors
		15.2.2. Monitor
		15.2.3. Control
		15.2.4. Intelligence and Logic
	15.3. Challenges and Limitations
		15.3.1. Interoperability
		15.3.2. Self-Management
		15.3.3. Maintainability
		15.3.4. Signalling
		15.3.5. Usability
		15.3.6. Power Aware/Efficient Consumption
		15.3.7. High Cost of Ownership
		15.3.8. Security and Privacy
		15.3.9. Acceptance and Reliability
		15.3.10. Calmness and Context-Awareness
		15.3.11. Architectural Readiness
	15.4. Conclusion
	References
16. Acceptance of Blockchain in Smart City Governance from the User Perspective
	16.1. Introduction
		16.1.1. Objectives
		16.1.2. Research Questions
		16.1.3. Chapter Organization
	16.2. Related Work
	16.3. Research Method
		16.3.1. Search Criteria
		16.3.2. Conceptual Model
			16.3.2.1. Social Cognitive Theory (SCT)
			16.3.2.2. Technology Acceptance Model (TAM)
			16.3.2.3. Urban Service Technology Acceptance Model (USTAM)
	16.4. Literature Review
		16.4.1. Prior Research on Technology Acceptance
		16.4.2. Factors and Hypotheses
		16.4.3. Factors Additional to USTAM
		16.4.4. Summary
	16.5. Proposed Future Methodology
		16.5.1. Approach
		16.5.2. Research Instrument
		16.5.3. Data Sample
		16.5.4. Data Analysis Methodology
	16.6. Discussion
	16.7. Conclusion
	References
17. Explainable AI in Machine/Deep Learning for Intrusion Detection in Intelligent Transportation Systems for Smart Cities
	17.1. Introduction
	17.2. Road ITS Technologies in Smart Cities
		17.2.1. Vehicular Area Network (VANET)
		17.2.2. VANET Communications
	17.3. Road ITS in Smart City Architecture
	17.4. Road ITS Security Issues and Challenges
		17.4.1. Road ITS Vulnerabilities
		17.4.2. Road ITS Threats and Attacks
	17.5. Intrusion Detection Against Road ITS Attacks
		17.5.1. Related Studies
		17.5.2. Discussion and Open Issues
	17.6. Explainable Artificial Intelligence in Cyber Security
		17.6.1. Explainable Artificial Intelligence Background Knowledge
		17.6.2. Related Studies in Explainable Artificial Intelligence for Cyber Security
		17.6.3. Discussion and Future Directions
	17.7. Case Study: DDoS Attacks in Road ITS and Their Impact on Smart Cities
		17.7.1. Road ITS and Smart City Simulation Environment in NS3
			17.7.1.1. Road ITS Simulation
			17.7.1.2. Smart Home Simulation
			17.7.1.3. Smart Hospital Simulation
			17.7.1.4. Telecom Network Simulation
			17.7.1.5. Other Infrastructures Simulations
		17.7.2. DDoS Attacks Modelling Scenarios
			17.7.2.1. Scenarios Targeting Road ITS Scenarios
			17.7.2.2. Scenarios Initiated from Road ITS
		17.7.3. Results, Analysis, and Discussion
		17.7.4. Road ITS-Smart City Security Impact Framework Evaluation
		17.7.5. Results and Discussion
	17.8. Conclusion
	References
18. Real-Time Identity Censorship of Videos to Enable Live Telecast Using NVIDIA Jetson Nano
	18.1. Introduction
	18.2. Background
	18.3. Materials
		18.3.1. Hardware
			18.3.1.1. NVIDIA Jetson Nano
			18.3.1.2. Camera
		18.3.2. Dataset
		18.3.3. Proposed Methodology
		18.3.4. Model Pipeline
	18.4. Experimental Study
		18.4.1. Data Augmentation
		18.4.2. Experiments
		18.4.3. Metrics
	18.5. Results
	18.6. Conclusions
	References
19. Smart Cities' Information Security: Deep Learning-Based Risk Management
	19.1. Introduction
	19.2. Literature Review
	19.3. Smart Cities' Security Based on DL
		19.3.1. Application of Association Rules
	19.4. Discussion
	19.5. Conclusion
	References
Index