Search

Browse Subject Areas

For Authors

Submit a Proposal

Tiny Machine Learning

Fundamentals, Applications and Security

Edited by Rajdeep Chakraborty, Rana Majumdar, S. Balamurugan and Sheng-Lung Peng
Series: Industry 5.0 Transformation Applications
Copyright: 2026   |   Expected Pub Date:2026/04/30
ISBN: 9781394347094  |  Hardcover  |  
522 pages

One Line Description
Stay at the forefront of the embedded AI revolution by mastering the specialized hardware and software strategies needed to bring high-performance machine learning to the world’s most resource-constrained devices.

Audience
Engineers, academics, researchers, and professionals in computer science, information technology, and electronics and communication.

Description
TinyML (tiny machine learning), short for tiny machine learning, represents a groundbreaking intersection of machine learning and embedded systems, enabling the deployment of intelligent applications on resource-constrained devices. It empowers these devices to perform complex tasks, like image and speech recognition, locally without relying on cloud servers. This burgeoning field opens up many possibilities, from enhancing IoT devices to revolutionizing healthcare and intelligent infrastructure. As technology advances, TinyML promises to make our everyday devices more innovative, responsive, and efficient than ever before. By bringing inference to resource-constrained hardware, TinyML supports real-time decision-making while addressing critical concerns such as latency, power consumption, and data privacy. This book presents an overview of TinyML, including its core principles, applications, challenges, and future directions. It meticulously explores the fundamentals of machine learning and deep learning, providing a solid foundation for understanding how these techniques are adapted for tiny devices. By delving into the hardware, software, and algorithms that specifically cater to TinyML, the book addresses the unique challenges of running machine-learning models on devices with limited processing power and memory. Featuring expert insights and real-world case studies, this volume is an essential guide to researchers and industry professionals looking for solutions for today’s resource-constrained devices.
Readers will find the volume:
• Delves into the burgeoning field of TinyML, where the power of machine learning is harnessed for resource-constrained devices;
• Serves as a comprehensive guide, equipping readers with the essential knowledge to develop and deploy TinyML applications;
• Explores the fundamentals of machine learning and deep learning, providing a solid foundation for understanding how these techniques are adapted for tiny devices;
• Introduces the hardware, software, and algorithms that specifically cater to TinyML, addressing the unique challenges of running machine-learning models on devices with limited processing power and memory.

Back to Top
Author / Editor Details
Rajdeep Chakraborty, PhD is a Professor in the Department of Computer Science and Engineering, Institute of Engineering and Technology, SAGE University, Indore, Madhya Pradesh, India with more than two decades of teaching experience and extensive involvement in research. He has made notable contributions through various publications, including patents, books, journal articles, and conference papers. His research interests include cryptography, network security, IoT, and blockchain.

Rana Majumdar, PhD is an Associate Professor at the Sister Nivedita University, Kolkata, West Bengal, India. He is the author of numerous research publications at the national and international levels, three books, one copyright, and 12 patents. His research focuses on machine learning, computer vision, software reliability engineering, digital image and video processing, and machine learning.

S. Balamurugan, PhD is the Director of Intelligent Research Consultancy Services, Coimbatore, Tamil Nadu, India. He has published more than 90 books, 300 articles in national and international journals and conferences, and 300 patents. He is a research consultant for many companies, startups, and micro-, small, and medium enterprises.

Sheng-Lung Peng, PhD is a Professor in the Department of Creative Technologies and Product Design and the Dean of the College of Innovative Design and Management at the National Taipei University of Business, Taiwan. In addition to his roles at NTUB, he holds honorary and adjunct professorships at several institutions and serves as the President of the Association of Taiwan Computer Programming Contest and the Association of Algorithms and Computation Theory. His research focuses on designing algorithms in artificial intelligence, bioinformatics, combinatorics, data mining, and networking, with more than 100 research papers published in these areas.

Back to Top

Table of Contents
Preface
Part I: Fundamentals
1. The Basics of TinyML: An Introductory Exploration
Darothi Sarkar and Monalisa Dey
1.1 Introduction to TinyML
1.2 Technological Underpinnings of TinyML
1.2.1 Core Concepts of Machine Learning in Tiny Devices
1.2.2 Differences between Traditional ML and TinyML
1.2.3 Power Efficiency and Resource Constraints
1.2.4 Role of IoT and Embedded Systems in TinyML
1.3 Real-World Applications of TinyML
1.3.1 Industrial IoT
1.3.2 Healthcare
1.3.3 Consumer Electronics
1.3.4 Agriculture
1.3.5 Autonomous Vehicles and Smart Sensors
1.4 Challenges and Limitations of TinyML
1.4.1 Computational and Memory Constraints
1.4.2 Security and Privacy Concerns
1.4.3 Deployment Challenges in Real-World Scenarios
1.4.4 Scalability Issues
Conclusion
References
2. Advances in TinyML: A Systematic Review of Architectures, Algorithms, and Innovations
Sumanta Chatterjee, Aritra Banerjee, Tania Biswas and Somya Ranjan Bhoi
2.1 Introduction
2.2 Background
2.2.1 Data and AI/ML
2.2.2 Edge Technology
2.2.2.1 Constraints of Cloud Computing
2.2.2.2 The Internet of Things
2.2.2.3 Hardware and Software Improvement
2.3 Tiny Machine Learning
2.3.1 Emergence and Breakthroughs
2.3.2 TinyML Over Other Technologies
2.4 TinyML Operations
2.4.1 Hardware: Infrastructures and Architectures
2.4.2 Software: Algorithms and Approach
2.4.2.1 Pruning
2.4.2.2 Quantization
2.4.2.3 Low-Rank Factorization
2.4.2.4 Huffman Coding
2.4.2.5 Knowledge Distillation
2.4.2.6 Tiny Neural Networks
2.4.3 Training and Deploying of Model
2.5 Application
2.5.1 IoT
2.5.1.1 Enabling IoT Devices with TinyML
2.5.1.2 Stand-Alone ML Services for Specific Tasks
2.5.1.3 Efficient Data Processing through Localized Models
2.5.1.4 Real-Time Data Control and Action
2.5.1.5 Alternative Data Analysis through Edge and Fog Computing
2.5.1.6 Low Latency with Edge and Fog Computing
2.5.1.7 Comparison of TinyML with Cloud, Fog and Edge Computing
2.5.1.8 TinyML Benchmarks and Their Role in Advancing ML
2.5.1.9 Benefits of TinyML Over Other Technologies
2.5.1.10 Challenges of Model Optimization for Low-Power Devices
2.5.2 Environmental Monitoring
2.5.2.1 Quality Control of Air
2.5.2.2 Monitoring Water Quality
2.5.2.3 Climate Change and Weather Monitoring
2.5.2.4 Wildlife and Ecosystem Monitoring
2.5.2.5 Hazard and Environmental Disaster Monitoring
2.5.3 Healthcare
2.5.3.1 Real-Time Health Monitoring
2.5.3.2 Early Disease Detection
2.5.3.3 Personalized Treatment
2.5.3.4 Wearable Health Devices
2.5.3.5 Remote Patient Monitoring
2.5.3.6 Cost-Efficient Healthcare
2.5.4 Agriculture
2.5.4.1 Precision Agriculture
2.5.4.2 Smart Irrigation
2.5.4.3 Pest and Disease Detection
2.5.4.4 Weather Monitoring
2.5.5 Security and Surveillance
2.5.5.1 Real-Time Threat Detection
2.5.5.2 Increased Privacy Due to On-Device Processing
2.5.5.3 Efficient Detection of Intruders
2.5.5.4 Smart Anomaly Detection
2.5.6 Industrial Automation
2.5.7 Finance
2.5.7.1 Real-Time Fraud Detection on the Edge
2.5.7.2 Personalized Financial Insights
2.5.7.3 Predictive Risk Management
2.5.7.4 On-Device Processing for Increased Security
2.5.8 Robotics
2.5.8.1 Intelligence on the Edge for Speedy Decisions
2.5.8.2 Energy-Efficient Processing of Data
2.5.8.3 Improved Smarter Automation and Control
2.5.8.4 Real-Time Monitoring and Alerting
2.5.8.5 Seamless Integration with Existing Devices
2.6 Challenges and Proposed Solutions
2.6.1 Limited Computational Power and Memory
2.6.2 Energy Efficiency
2.6.3 Model Optimization and Compression
2.6.4 Scalability to Domains
2.6.5 Real-Time Processing and Latency
2.6.6 Data Privacy and Security
2.6.7 Interoperability and Standards
2.6.8 Data Availability and Labeling
2.6.9 Hardware Development and Cost
2.6.10 Continuous Learning and Model Update
2.7 Impacts of TinyML
2.8 Sustainable Development Through TinyML
2.8.1 Decreased Latency
2.8.2 Offline Functionality
2.8.3 Strengthened Security and Privacy
2.8.4 Power Efficiency
2.8.5 Cost Effectiveness
2.9 Conclusion
2.10 Future Scope
References
3. Edge Intelligence and Trust: The Synergy of TinyML, IoT, and Blockchain in Modern Applications
Abhishek Bhattacharya, Soumi Dutta, Anupam Ghosh, Arijit Dutta, Prabuddha Chatterjee and Sangeeta Banik
3.1 Introduction
3.1.1 TinyML
3.1.2 Internet of Things (IoT)
3.1.3 Blockchain
3.2 Literature Review
3.2.1 TinyML Literature Review
3.2.2 IoT Literature Review
3.2.3 Blockchain Literature Review
3.3 Applications
3.3.1 Smart Home Automation and Security Using IoT
3.3.2 Wildlife Monitoring and Conservation Using IoT
3.3.3 Smart Cities and Traffic Management Using IoT
3.3.4 Fraud Detection in Financial Transactions Using Blockchain
3.3.5 Healthcare Data Integrity and AI-Driven Diagnostics Using Blockchain
3.3.6 Energy Management in Smart Grids Using Blockchain
3.4 Discussion
3.5 Conclusion
References
4. Use Cases of TinyML
S. Sharmila Devi
4.1 Introduction
4.2 Use Cases of TinyML
4.2.1 Visual Analysis
4.2.2 Gesture Detection
4.2.3 Facial Recognition
4.2.4 Abnormality Identification
4.2.5 Exceptional and Environmental Preservation
4.2.6 Autonomous Driving and Traffic Coordination
4.2.7 Posture Assessment
4.2.8 Identification of Respiratory Symptoms Linked to Coughing
4.2.9 Speech Sound Identification
4.2.10 Pre-Screening for Oral Tongue Abnormalities
4.2.11 Intelligent Sensors for Residential Automation
4.2.12 Voice Assistants for On-Site Command Execution
4.3 Conclusion
4.4 Future Scope
References
Part II: Applications
5. Advancing Smart Devices and IoT: Research Insights and Directions
Ajay Verma, Nahida Majeed Wani and Girraj Kumar Verma
5.1 Introduction
5.1.1 What is a Smart Device?
5.1.2 Objectives of the Chapter
5.2 How Smart Devices Work
5.2.1 Data Collection
5.2.2 Data Processing
5.2.3 Decision Making
5.2.4 Communication and Control
5.3 The Need for Smart Devices in the Real World
5.3.1 Home Automation
5.3.2 Healthcare
5.3.3 Energy Management
5.3.4 Entertainment
5.3.5 Efficiency at Work
5.4 Properties of Smart Devices
5.4.1 Connection
5.4.2 Being Aware of the Context
5.4.3 Computing on Its Own
5.4.4 User Interaction
5.4.5 Interoperability
5.4.6 Efficiency of Energy
5.4.7 Safety and Privacy
5.4.8 Scalability
5.4.9 Processing in Real Time
5.4.10 Adaptability and Learning
5.4.11 Mobility
5.4.12 Interaction with the Environment
5.5 Connection to the Internet of Things (IoT)
5.6 Security and Privacy: Keeping the IoT Landscape Safe
5.6.1 Real-Life Risks to IoT Security and Privacy
5.6.2 Ways to Make IoT Security Better
5.6.3 Ways to Protect Your Privacy
5.6.3.1 Privacy by Design
5.6.3.2 Anonymizing Data
5.6.3.3 Control by the User
5.6.3.4 Legislative Frameworks
5.7 Trends and Research Opportunities in the Future
5.8 Conclusion
References
6. TinyML for Smart Devices and IoT: Enabling Efficient and Intelligent Applications
Neeta A. Ukirade
6.1 Introduction
6.2 Tools and Frameworks for TinyML Development
6.3 Key Techniques in TinyML for IoT
6.4 Applications of TinyML in Smart IoT Devices
6.4.1 Healthcare
6.4.2 Automotive
6.4.3 Agriculture
6.4.4 Security and Surveillance
6.4.5 Industry
6.5 Challenges and Limitations
6.5.1 Low Power
6.5.2 Limited Memory
6.5.3 Lack of Benchmarking Tools
6.5.4 Heterogeneity
6.5.5 Data and Network Management
6.5.6 Latency
6.6 Future Directions
6.7 Conclusion
References
7. Predictive Maintenance Using Tiny Machine Learning: A Revolutionary Approach to Proactive Equipment Maintenance
G. JayaLakshmi, Ch. JayaLakshmi and M. Ramesh
7.1 Introduction
7.2 Predictive Maintenance: The Need for Proactivity
7.2.1 Limitations
7.2.2 Benefits of Predictive Maintenance
7.3 TinyML: Scope, Advantages, and Applications
7.3.1 Key Advantages of TinyML
7.3.2 Examples of TinyML Applications in Resource-Constrained Environments
7.4 TinyML in Predictive Maintenance: Key Components and Implementation Framework
7.5 Analyzing Real-World Applications and Case Studies of TinyML-Based Predictive Maintenance Systems
7.6 Conclusion
7.7 Future Scope
References
8. TinyML and IoT in Agriculture: Boosting Real-Time Efficiency, Autonomy, and Resilience in Smart Farming
Shanthalakshmi M., Deepika N., Avvudaiyappan R.M. and Prince Raj J.
8.1 Introduction
8.2 Smart Fertilizer Distribution Using Soil and Crop Data
8.3 Weed Detection Using TinyML and IoT
8.4 Animal Intrusion Detection in Crops
8.5 Disease Prevention and Detection
8.6 Enhancing Smart Irrigation with TinyML for Climate Prediction and Optimization Existing Systems
8.6.1 Drawbacks of Existing Systems
8.6.2 The Idea: TinyML for Climate Prediction and Optimization
8.6.3 Benefits of TinyML in Irrigation
8.6.4 How It Solves the Drawbacks
8.6.5 Efficiency and Overall Impact
8.7 Conclusion
8.8 Future Scope
References
9. TinyML and IoT for Predictive Maintenance and Real-Time Decision Support in Automotive Air Conditioning
G. Bhavani and C. Jeyalakshmi
9.1 Introduction
9.1.1 TinyML-IoT
9.1.2 Predictive Maintenance and Real-Time Decision Making
9.2 Architecture Overview
9.2.1 Working Principle of Vehicle AC System
9.2.2 Sensors Used to Collect Real-Time Data
9.2.3 Placeholders for Data Monitoring Sensors
9.3 Technologies Enabling TinyML
9.3.1 TinyML Frameworks for Automotive AC Monitoring
9.3.1.1 TensorFlow Lite for Microcontrollers (TFLM)
9.3.1.2 Edge Impulse for TinyML and IoT Applications
9.3.2 Edge Device Hardware
9.3.3 Sensor Technologies
9.3.4 Communication Protocols
9.3.5 Data Pre-Processing Techniques
9.3.6 TinyML Model Optimization
9.3.7 Deployment and Monitoring
9.4 Advantages of a Properly Functioning AC System
9.5 Advantages of TinyML in Real-Time Data Monitoring
9.6 Challenges
9.7 Conclusion
9.8 Future Scope
References
10. Automated Harm Detection: Enhancing Women’s Safety in Real Time
Shoban S., Rohith V., Shanthalakshmi M. and Deepika N.
10.1 Introduction
10.1.1 Challenges in Existing Wearable Models
10.1.2 TinyML for Local Analysis
10.2 Prior Knowledge
10.2.1 Psychophysiology
10.2.2 Heart Rate Variability
10.3 Related Works
10.3.1 Wrist Vascular Biometric Recognition
10.3.2 Stress Prediction Model
10.3.3 Wearable Healthcare Devices - Material Advancement
10.3.4 Today’s Invasive Continuous Health Monitoring
10.4 Proposed Methodology
10.4.1 Wrist Authorization Using Vein Pattern Detection
10.4.1.1 About It
10.4.1.2 Workflow
10.4.1.3 Why This
10.4.2 Health Monitoring
10.4.2.1 Significant Factors for Health Monitoring
10.4.2.2 Workflow
10.4.3 Location Data Analysis
10.4.3.1 Data Gathering
10.4.3.2 Model Selection and Workflow
10.4.3.3 Why Temporal-Spatial Model
10.4.4 Integration Part
10.4.4.1 Total Workflow
10.4.4.2 Pseudo Code
10.4.4.3 Dependent Sensor Activation (Optimal Memory and Power Utilization)
10.4.4.4 Why Dependent Sensor Activation?
10.5 Challenges
10.6 Future Scope
References
11. Butterfly Optimization with Random Forest for COVID-19 Prediction Using Lung Image
Sivanantham Kalimuthu, Ramkumar N., Arun Prakash N., Boorneush M. and Dhusiyanth M.
11.1 Introduction
11.1.1 COVID-19
11.1.2 COVID-19 Nomenclature
11.1.3 COVID-19 Causes
11.1.4 Pre-Processing
11.1.5 Butterfly Optimization Algorithm (BOA)
11.1.6 Random Forest (RF)
11.2 Literature Survey
11.3 Proposed Research Methodology
11.3.1 Lung Nodules
11.3.2 Pre-Processing
11.3.3 Median Filter
11.3.4 Butterfly Optimization Algorithm (BOA)
11.3.5 Random Forest Algorithm (RF)
11.3.6 Feature Extraction
11.3.7 Image Classification
11.4 Implementation Results
11.4.1 Input Image
11.4.2 Feature Extraction
11.4.3 Performance Evaluation
11.4.4 F-Measure
11.4.5 Recall
11.4.6 Accuracy
11.4.7 Precision
11.5 Conclusion
11.6 Future Scope
Bibliography
Part III: Security
12. AI-Powered Resilience and Privacy Preservation in Cloud-IoT Environments for Smart Devices Using Fog Computing Methodologies
Biplab Gope and Soumen Santra
12.1 Introduction
12.1.1 Challenges in Resilience and Privacy in Cloud-IoT Ecosystems
12.1.2 The Role of AI in Addressing Resilience and Privacy Challenges
12.1.3 Objectives and Structure of the Chapter
12.2 AI-Powered Resilience Mechanisms
12.2.1 Fault Prediction
12.2.2 Dynamic Task Migration
12.3 Privacy Preservation Techniques
12.3.1 Federated Learning
12.3.2 Homomorphic Encryption
12.3.3 Differential Privacy
12.4 Fog Computing as an Enabler
12.5 Key Concepts
12.5.1 Resilience in IoT Systems
12.5.2 Privacy in Cloud-IoT Environments
12.5.3 Fog Computing
12.5.4 AI in IoT Systems
12.6 Results
12.7 Current Practices and Challenges
12.7.1 Resilience Enhancement
12.7.2 Privacy Preservation
12.8 Challenges
12.8.1 Scalability Issues
12.8.2 Heterogeneous Ecosystems
12.8.3 Energy Consumption
12.9 Proposed Solutions
12.9.1 Resilience through AI in Fog Computing
12.9.1.1 Fault Prediction Models
12.9.1.2 Autonomous Recovery Systems
12.9.1.3 Real-Time Monitoring
12.9.2 Privacy Preservation through AI and Fog
12.9.2.1 Federated Learning for Privacy
12.9.2.2 Differential Privacy in Data Sharing
12.9.2.3 Secure Data Storage and Transmission
12.10 Applications and Use Cases
12.10.1 Smart Healthcare
12.10.2 Smart Cities
12.10.3 Industrial-IoT (I-IoT)
12.10.4 Smart Homes
12.11 Technological Frameworks
12.11.1 Architecture of AI-Fog-IoT Systems
12.11.2 Implementation Stack
12.12 Conclusion
12.13 Future Scope
12.13.1 Enhanced AI Algorithms for Fog Nodes
12.13.2 Advanced Privacy Preservation Techniques
12.13.3 Increased Adoption in Emerging Technologies
12.13.4 Integration with Blockchain
12.13.5 Emergence of 6G Networks
12.13.6 Green Computing and Energy Efficiency
12.13.7 Collaboration with Quantum Computing
12.13.8 Autonomous Fog Knots and Self-Healing System
12.13.9 Human-Focused AI and Interpretable AI (XAI) in Fog Data
12.13.10 AI-Powered Cybersecurity in Fog Environments
12.13.11 Next-Gen Secure Data Sharing Mechanisms
12.13.12 AI-Augmented Digital Twins for Fog Computing
12.13.13 Multi-Layer Hybrid AI for Adaptive Fog-IoT Networks
References
13. Data Privacy and Transmission Security
Saptarshi Kumar Sarkar, Anupama Sen and Piyal Roy
13.1 Introduction
13.1.1 Overview of Data Privacy and Transmission Security
13.1.2 Importance in the Modern Digital Landscape
13.1.3 Objectives
13.2 Foundations of Data Privacy
13.2.1 Definition and Core Concepts
13.2.2 Principles of Data Minimization and Consent
13.2.3 Control and Ownership of Personal Data
13.2.4 Legal Frameworks and Regulations
13.2.5 Global Trends in Data Privacy Regulations
13.3 Transmission Security
13.3.1 Importance of Securing Data During Transmission
13.3.2 Overview of Cryptography and Encryption
13.3.3 Digital Signatures and Authentication Methods
13.3.4 Secure Communication Protocols (SSL, TLS)
13.3.5 End-to-End Encryption: Ensuring Confidentiality and Integrity
13.4 Emerging Threats to Data Privacy and Transmission Security
13.4.1 Increasing Complexity of Cyber Threats
13.4.2 Types of Cyber Attacks (Ransomware, Phishing, MITM)
13.4.3 Vulnerabilities in Transmission and Storage
13.4.4 Ethical Considerations: Balancing Security and Privacy
13.5 Impact of Emerging Technologies
13.5.1 Quantum Computing and Its Implications on Encryption
13.5.2 Artificial Intelligence (AI) and Data Privacy Challenges
13.5.3 Internet of Things (IoT) and Transmission Security
13.5.4 Blockchain Technology: Enhancing or Threatening Data Privacy
13.6 Challenges in Ensuring Data Privacy and Secure Transmission
13.6.1 Managing the Evolving Nature of Cyber Threats
13.6.2 Technological Limitations and Vulnerabilities
13.6.3 Legal and Regulatory Gaps
13.6.4 Ethical Dilemmas: Over-Surveillance and Data Sovereignty
13.7 Practical Approaches to Enhancing Data Privacy and Transmission Security
13.7.1 Risk Mitigation Strategies for Individuals and Organizations
13.7.2 Encryption Best Practices and Security Protocols
13.7.3 Legal and Ethical Guidelines for Data Protection
13.7.4 Integrating Technical, Legal, and Ethical Solutions
13.8 Conclusion
References
14. Security and Privacy Concerns for Blockchain-Enabled Federated Learning
Partha Ghosh, Ananya Biswas, Suradhuni Ghosh, Rima Bhowmik and Ankita Barua
14.1 Introduction
14.1.1 Security and Privacy Concerns in Blockchain Enabled FL
14.1.2 Combination of Blockchain and Federated Learning
14.2 Importance of Security and Privacy
14.3 Architecture of Federated Learning
14.4 Difference between Centralized Learning, Distributed Learning, and Federated Learning
14.5 Sources of Vulnerabilities in Federated Learning
14.5.1 Communication
14.5.2 Gradient Leakage
14.5.3 Compromised Clients
14.5.4 Compromised Server
14.5.5 Aggregation Algorithm
14.5.6 Non-Malicious Failure
14.5.7 Distributed Nature of FL
14.5.8 Federated Learning Environment Scope
14.5.9 Model Deployment
14.6 Security Threats in Federated Learning
14.6.1 Poisoning Attack
14.6.2 Inference Attack
14.6.3 Communication Attack
14.6.4 Free-Riding Attacks
14.7 Defense Mechanism in Federated Learning System
14.7.1 Differential Privacy
14.7.2 Secure Multi-Party Computation
14.7.3 Anomaly Detection
14.7.4 Robust Aggregation
14.7.5 Federated Distillation
14.7.6 Trusted Execution Environment (TEE)
14.7.7 Pruning
14.7.8 Zero-Knowledge Proofs (ZKP)
14.7.9 Adversarial Training
14.7.10 Federated Multi-Task Learning (FML)
14.7.11 Moving Target Defense (MTD)
14.7.12 Recognizing Legitimate Clients
14.8 Federated Learning Schemes
14.8.1 Local Learning Model
14.9 Federated Learning: An Approach to Healthcare in IIoE that Protects Privacy
14.9.1 Benefits for Patient Privacy in IIoE
14.9.2 FL is Being Investigated for a Number of IIoE Healthcare Uses
14.10 Homomorphic Encryption (HE) Method in IIoE-Focused Federated Learning
14.10.1 Challenges of Homomorphic Encryption in FL for IIoE
14.10.2 A Short Case Study of the Usage of Homomorphic Encryption in FL for IIoE
14.11 Blockchain-Powered Federated Learning
14.11.1 An Explanation of Each Entity’s Specific Functions in the FBS
14.11.2 Data Preprocessing and Distribution for FL
14.12 Decentralized Data Sharing in Healthcare
14.12.1 Structure of Blockchain
14.12.2 Steps for Decentralized Data Sharing
14.12.3 Verifying the Robustness of the System
14.13 Public Key Infrastructure (PKI) for the System
14.14 Protecting Privacy with Cross-Chained Fl Techniques
14.14.1 Outline of Cross-Chained FL
14.14.2 Resolution for Privacy Conservation
14.15 Use of Blockchain-Enabled FL to Preserve Privacy
14.15.1 Healthcare
14.15.2 Industry 5.0
14.15.3 Internet of Vehicles (IoV)
14.16 Challenges and Solutions
14.16.1 Effective Data Management
14.16.2 Heterogeneity and Interoperability
14.16.3 Server-Side Attack
14.16.4 Optimization
14.16.5 Inference Attack
14.16.6 Client-Side Attack
14.16.7 Data Leakage
14.16.8 Data Privacy
14.16.9 Communication Overhead
14.16.10 Data Sharing in Horizontal Federated Learning
14.16.11 Anomaly Detection
14.16.12 Data Integrity
14.16.13 Data and Model Attack
14.17 Open Research Challenges
14.17.1 Sparsification Enabled Federated Learning
14.17.2 Data Heterogeneity Aware Clustering-Enabled Federated Learning
14.17.3 Mobility-Aware Federated Learning
14.17.4 Homomorphic Encryption-Enabled FL
14.17.5 Protected and Trustful Combination Enabled FL
14.17.6 Interference-Aware Resource Efficient Federated Learning
14.17.7 Quantization Enabled Federated Learning
14.17.8 Adaptive-Resource-Sharing Enabled FL
14.18 Conclusion and Future Direction
Acknowledgment
References
15. Adversarial Attacks and Defenses in Security
Sudeshna Dey, Siddhartha Chatterjee, Sumita Gupta and Sima Das
15.1 Introduction
15.2 Fundamentals of Federated Learning
15.2.1 Types of Federated Learning
15.2.1.1 Horizontal Federated Learning (HFL)
15.2.1.2 Vertically Federated Learning (VFL)
15.2.1.3 Federated Transfer Learning (FTL)
15.2.1.4 Cross-Silo FL
15.2.1.5 Cross-Device FL
15.3 Security and Privacy Threats in FL
15.3.1 Creating the First Model
15.3.2 Model Training
15.3.3 Update the Global Model
15.4 Attacks in Federated Learning
15.4.1 Poisoning Attacks and Defenses
15.4.1.1 Data Poisoning Attacks and Defenses
15.4.1.2 Model Poisoning Attacks and Defenses
15.4.2 Inference Attacks and Defenses
15.4.2.1 Secure Multiparty Computation (SMC)
15.4.2.2 Differential Privacy (DP)
15.4.2.3 Data Anonymization
15.4.2.4 Homomorphic Encryption
15.4.2.5 Trusted Execution Environment (TEE)
15.4.2.6 Blockchain
15.5 Problems and Committing Directions
15.5.1 Contribution Assessments and Compensation Strategies
15.5.2 Emphasize on Solidity or Privacy
15.5.3 Impact Metrics and Incentive Mechanisms
15.5.4 Trade-Off in Safety, Communication and Computing
15.5.5 Alternative Meaning of Methods of Attack
15.5.6 Applications for FL Systems
15.6 Conclusion
References
16. Ethical and Technical Foundations of Privacy-Preserving Federated Learning
Muhammad Rifthy Kalideen
16.1 Introduction
16.2 Foundations of Federated Learning
16.2.1 Overview of Federated Learning
16.2.1.1 Comparing Federated Learning and Traditional Centralized Learning
Approaches
16.2.1.2 Privacy Concerns in FL
16.3 Ethical Foundations of Privacy in Federated Learning
16.3.1 Ethical Principles in Data Privacy
16.3.2 Ethical Dilemmas in FL
16.3.3 Regulatory and Legal Considerations
16.3.4 Societal Impact and Public Perception
16.4 Technical Foundations of Privacy-Preserving Federated Learning
16.4.1 Privacy-Preserving Techniques
16.4.2 Differential Privacy (DP)
16.4.3 Secure Multi-Party Computation
16.4.4 Homomorphic Encryption
16.4.5 Federated Learning with Trusted Execution Environments
16.4.6 Architecture of Privacy-Preserving FL Systems
16.4.7 Evaluation of Privacy-Preserving Mechanisms
16.5 Interplay Between Ethical and Technical Foundations
16.5.1 Ethical Implications of Technical Choices
16.5.2 Technical Solutions to Ethical Challenges
16.5.3 The Need for Ethical Guidelines in Technical Implementation
16.6 Case Studies and Real-World Applications
16.6.1 Health
16.6.2 Finance
16.6.3 Cross-Industry Insights
16.7 Future Directions and Emerging Trends
16.7.1 Advances in Privacy-Preserving Techniques
16.7.2 Quantum Federated Learning
16.7.3 Ethical Challenges on the Horizon
16.7.4 The Role of AI Ethics in FL
16.8 Conclusion
References
17. Integrating Security Measures in CLAHE-Enhanced YOLOV8 Model for Underwater Object Detection
Niyati Sahoo, Sanjukta Mohanty and Arup Abhinna Acharya
17.1 Introduction
17.2 Background
17.2.1 Yolo
17.2.2 Different YOLOv8 Types
17.2.3 Blocks Used in YOLOv8 Architecture
17.2.3.1 Convolutional Block (Conv Block)
17.2.3.2 Bottleneck Block
17.2.3.3 C2f Block
17.2.3.4 Spatial Pyramid Pooling Fast (SPPF) Block
17.2.3.5 Detect Block
17.2.4 Architecture of YOLOv8
17.2.4.1 Backbone Section
17.2.4.2 Neck and Head Section
17.2.5 Image Enhancement
17.3 Related Works
17.3.1 Literature Review for Underwater Object Detection Using YOLO Technique
17.3.2 Literature Review on Image Enhancement
17.4 Proposed Approach
17.4.1 Data Collection
17.4.2 Enhancing Underwater Images
17.4.3 Performance Evaluation Metrics
17.5 Experiment and Results
17.6 Frequently Occurring Threats and Mitigation Policy
17.6.1 Data Integrity and Tampering
17.6.2 Adversarial Attacks
17.6.3 Sensor Spoofing and False Target Generation
17.6.4 Denial of Service (DoS) Attacks
17.6.5 Privacy and Data Privacy Concerns
17.6.6 Model Inversion and Reverse Engineering
17.6.7 Physical Attacks on Hardware
17.7 Conclusion
17.8 Future Scope
References
18. TinyML Deployment for Resource-Constrained Devices in IoT Applications with Attribute-Based Encryption Scheme
R. Lavanya and V. Thanigaivelan
18.1 Introduction
18.2 Resource-Constrained Devices
18.2.1 Particular to a Model Risk
18.3 Background for Attribute-Based Encryption
18.3.1 Types of ABE
18.4 Related Work in ABE and Other Security Schemes
18.5 Local Interpretable Model-Agnostic Explanations
18.6 Conclusion
References
19. Deep Learning-Based Adversarial Attack Detection for Cloud-IoT Systems
Amit Kumar, Sachin Ahuja and Ganesh Gupta
19.1 Introduction
19.2 Background and Motivation
19.2.1 Cloud-IoT Systems
19.2.2 Adversarial Attacks in IoT
19.2.2.1 Types of Adversarial Attacks
19.2.2.2 Impacts of Adversarial Attacks
19.3 Deep Learning for Intrusion Detection in Cloud-IoT Systems
19.3.1 Existing Techniques
19.3.2 Deep Learning’s Role in Security
19.3.2.1 CNN-Based Detection Systems
19.3.3 RNN and LSTM for Temporal Anomaly Detection
19.3.4 Generative Adversarial Networks (GANs) for Attack Detection
19.4 Case Study: Adversarial Attack Detection in Smart Grid Systems
19.4.1 Overview of Smart Grids in India
19.4.2 Detecting Adversarial Attacks in Energy Consumption Data
19.4.2.1 Results and Analysis
19.5 Challenges and Future Directions
19.5.1 Challenges in Detecting Adversarial Attacks
19.5.2 Future Directions
19.6 Conclusion
Bibliography
Index

Back to Top



Description
Author/Editor Details
Table of Contents
Bookmark this page