Adversarial Machine Learning: Mechanisms, Vulnerabilities, And Strategies For Trustworthy Ai

Adversarial Machine Learning is a practical and comprehensive guide to securing AI systems against modern adversarial threats. Covering the full machine learning lifecycle, the book explains attacks such as evasion, poisoning, backdoor, privacy, and generative AI exploits, while also exploring defence strategies including adversarial training, certified defences, and privacy-preserving techniques. Designed for developers, security professionals, and AI leaders, it provides both technical depth and real-world guidance for building resilient and trustworthy AI systems.