Adversarial Learning and Secure AI

This textbook is the first on adversarial learning and introduces methods for defending against attacks and making AI more robust. It includes feasible hands-on student projects and covers deep neural networks, logistic regression, naïve Bayes classifiers, and support vector machines. It is written for senior undergraduate and first-year graduate courses and offers online resources for instructors and students.

Format: Hardback
Length: 350 pages
Publication date: 31 August 2023
Publisher: Cambridge University Press

This groundbreaking textbook is the first to delve into the realm of adversarial learning, offering a comprehensive framework for student learning. It begins by introducing the vulnerabilities of deep learning, highlighting the potential risks and challenges associated with these powerful algorithms. Subsequently, it presents robust methods for defending against attacks, ensuring the security and reliability of AI systems. To facilitate a strong connection between theory and practice, the book elucidates and evaluates attack-and-defense scenarios alongside real-world examples, providing students with a practical understanding of the field.

To enhance the learning experience, the book includes a series of feasible, hands-on student projects that gradually increase in difficulty. These projects not only reinforce the theoretical concepts but also develop students' Python and PyTorch skills, making them well-equipped for real-world applications. Each chapter concludes with thought-provoking questions that encourage classroom discussions and promote active learning.

In addition to deep neural networks, the textbook delves into other important machine learning algorithms such as logistic regression, naïve Bayes classifiers, and support vector machines. This broad coverage enables students to gain a comprehensive understanding of the field and prepares them for various research and industry roles.

Written for senior undergraduate and first-year graduate courses, this textbook serves as a valuable window into research methods and current challenges in the field. It provides a solid foundation for students interested in pursuing advanced studies in adversarial learning and AI security.

To support instructors, the book includes comprehensive lecture slides and image files, which can be used to enhance classroom presentations. Additionally, software for early course projects is available, allowing students to apply the concepts learned in the book and develop practical skills.

In summary, this textbook is an essential resource for students, researchers, and practitioners seeking to excel in the field of adversarial learning and AI security. Its comprehensive coverage, practical projects, and engaging teaching approach make it an invaluable tool for anyone looking to advance their knowledge and skills in this rapidly evolving domain.

Weight: 866g
Dimension: 175 x 251 x 25 (mm)
ISBN-13: 9781009315678