Название: AI for Cybersecurity: Research and Practice
Автор: Houbing Herbert Song, Elisa Bertino, Alvaro Vasquez, Huihui Helen Wang, Yan Shoshitaishvili, Sumit Kumar Jha
Издательство: Wiley-IEEE Press
Год: 2026
Страниц: 659
Язык: английский
Формат: pdf (true), epub
Размер: 31.5 MB

Informative reference on the state of the art in cybersecurity and how to achieve a more secure cyberspace.

AI for Cybersecurity presents the state of the art and practice in AI for cybersecurity with a focus on four interrelated defensive capabilities of deter, protect, detect, and respond. The book examines the fundamentals of AI for cybersecurity as a multidisciplinary subject, describes how to design, build, and operate AI technologies and strategies to achieve a more secure cyberspace, and provides why-what-how of each AI technique-cybersecurity task pair to enable researchers and practitioners to make contributions to the field of AI for cybersecurity.

Written by a team of highly qualified experts in the field,AI for Cybersecurity discusses topics including:
Robustness and risks of the methods covered, including adversarial ML threats in model training, deployment, and reuse
Privacy risks including model inversion, membership inference, attribute inference, re-identification, and deanonymization
Forensic and formal methods for analyzing, auditing, and verifying security- and privacy-related aspects of AI components
Use of generative AI systems for improving security and the risks of generative AI systems to security
Transparency and interpretability/explainability of models and algorithms and associated issues of fairness and bias

The application of Data Science to cybersecurity can be approached from two main perspectives: (1) defensive, where data science techniques support the development of protection and prevention systems, such as malware detection and network attack identification; and (2) offensive, where data analysis is used to design and execute more sophisticated cyberattacks. This chapter focuses on the defensive perspective, exploring how data science and AI are leveraged to create more effective security systems. The intersection of data science, AI, and ML is transforming cybersecurity by enabling the rapid analysis of massive data volumes, identifying patterns, and making accurate predictions far faster than humans can. These technologies are essential for detecting and mitigating cyber threats in real time, ultimately strengthening organizational security. For instance, AI can identify anomalous patterns that signal early stage cyberattacks, which is particularly critical for countering advanced techniques such as fileless malware. Moreover, AI facilitates the development of automated incident response systems and predictive tools. It can even simulate attack scenarios to train professionals, fostering a proactive and resilient cybersecurity culture.

This edited book, AI for Cybersecurity: Research and Practice, aims to present the state of the art and the state of the practice of AI for cybersecurity with a focus on four interrelated defensive capabilities (deter, protect, detect, and respond). Toward this goal, this book is organized into two parts: Research and Practice.

Part 1 is composed of twelve chapters. This part presents the state of the art of AI for cybersecurity, including large language models (LLMs) (Chapter 1), privacy leakage detection (Chapter 2), intrusion detection (Chapter 3), generative AI (Chapter 4), threat detection (Chapter 5), privacy-preserving (Chapter 6), federated learning (Chapter 7), communications and network security (Chapter 8), secure by design (Chapter 9), Internet of Things (IoT) security (Chapter 10), object tracking (Chapter 14), and physical layer security (Chapter 17).

Part 2 is composed of eight chapters. This part presents the state of the practice of AI for cybersecurity, including distributed denial of service (DDoS) attacks (Chapter 11), education (Chapter 12), ethics (Chapter 13), malware (Chapter 15), supply chain vulnerabilities (Chapter 16), smart grid (Chapter 18), geopolitical dimensions (Chapter 19), and nuclear nonproliferation (Chapter 20).

This book is primarily intended for use by anyone who aspires to make contributions to AI for cybersecurity through research or practice. On the one hand, the state of the art in research in AI for cybersecurity will enable the readers to grow into successful researchers in AI for cybersecurity; on the other hand, the practice in a variety of application domains, including education, energy, and defense, will enable the readers to be successful in applying AI to enhance cybersecurity in practice. This book envisions a greater community committed to advancing research and practice at the confluence of cybersecurity and AI, and to transitioning its findings into engineering practice.

Contents: