A major challenge for machine learning solutions is that their efficiency in real-world applications is constrained by the current lack of ability of the machine to explain its decisions and activities to human users. Biases based on race, gender, age or location have been a long-standing risk in training AI models. Furthermore, AI model performance can degrade because production data differs from training data.

• Chapter 1: Explaining AI for Safeguarding and Securing Internet of Things (IoT) Systems - An Introduction
• Chapter 2: Securing the Internet of Things: Architectures and Designs
• Chapter 3: Convergence of Internet of Things and Computing Technologies
• Chapter 4: Security Vulnerabilities in Internet of Things: Attack Surfaces, Threats, and Defense
• Chapter 5: Black-Box Machine Learning for IoT Security
• Chapter 6: Explainable Intelligence for safeguarding IoT
• Chapter 7: Explainability Methods in Explainable Security Intelligence: Fine-grained Taxonomy
• Chapter 8: Intrinsically Explainable Security Intelligence
• Chapter 9: Model-Agnostic Methods for Globally Interpretable Machine Learning
• Chapter 10: Model-Agnostic Methods for Locally Explainable AI to Secure IoT system
• Chapter 11: Explainability Evaluation Metrics for Explainable Security Intelligence in the Internet of Things
• Chapter 12: Adversarial Attacks and Defense in Explainable Security Intelligence
• Chapter 13: Federated Learning Meets Explainable AI at the Edge of Things
• Chapter 14: Explainable Security Intelligence for Zero-Trust IoT
• Chapter 15: Explainable Security Intelligence in IoT Applications