Adversarial Robustness for Machine Learning

Adversarial Robustness for Machine Learning summarizes recent progress in this area and introduces popular algorithms for adversarial attack, defense, and verification. It focuses on image classification applications, discusses adversarial examples beyond image classification, other threat models, and applications on adversarial robustness. The book is a thorough literature review and can be used as a textbook for graduate courses on adversarial robustness or trustworthy machine learning.

Format: Paperback / softback
Length: 298 pages
Publication date: 01 September 2022
Publisher: Elsevier Science Publishing Co Inc

Adversarial Robustness for Machine Learning is a comprehensive and up-to-date book that provides a comprehensive overview of the recent advancements in this field. It summarizes the state-of-the-art techniques and algorithms for adversarial attack, defense, and verification, with a particular focus on image classification applications, which are widely regarded as the standard benchmark in the adversarial robustness community.

The book is organized into several sections, each addressing different aspects of adversarial robustness. The first section introduces the fundamentals of adversarial attacks and their potential threats to machine learning models. It covers various attack techniques, such as input perturbation, gradient manipulation, and model poisoning, and explains how they can be used to fool ML algorithms.

The second section focuses on verification methods, which aim to assess the robustness of ML models against adversarial attacks. It discusses popular verification algorithms, such as adversarial training, robustness testing, and defensive distillation, and their effectiveness in detecting and mitigating adversarial examples.

The third section explores defense techniques, which aim to improve the robustness of ML models against adversarial attacks. It discusses various defense strategies, such as regularization, weight initialization, adversarial training, and transfer learning, and their ability to enhance the generalization and robustness of ML models.

In addition to these sections, the book also discusses adversarial examples beyond image classification, other threat models beyond testing time attack, and applications on adversarial robustness. It provides a thorough literature review that summarizes the latest progress in the area, making it an invaluable resource for researchers seeking to conduct future research in this field.

Furthermore, the book can also serve as a textbook for graduate courses on adversarial robustness or trustworthy machine learning. It covers the theoretical foundations, practical implementations, and applications of adversarial robustness, providing a comprehensive understanding of this important topic.

Overall, Adversarial Robustness for Machine Learning is a must-read for anyone interested in machine learning, security, and artificial intelligence. It provides a comprehensive and up-to-date overview of the latest advancements in adversarial robustness, and offers valuable insights and techniques for practitioners and researchers alike.

Weight: 474g
Dimension: 153 x 228 x 19 (mm)
ISBN-13: 9780128240205