Summary:

Machine learning (ML) and artificial intelligence (AI) algorithms are nowadays ubiquitous for image analysis and signal processing tasks. Likewise, the autonomous and robotic systems deployed in military and Defence applications, such as security, surveillance and targeting systems, are or will be based on ML/AI algorithms. Despite their high efficiency and accuracy, ML and AI algorithms can be fooled by an adversary through manipulation or spoofing of data (also known as adversarial attacks), which poses serious threat for security sensitive applications.   

The integration of quantum computing in ML and AI is rapidly progressing, which offers an opportunity to create new quantum ML/AI models which are designed to fundamentally exploit quantum mechanical properties for gaining advantages such as in training speed or feature extraction accuracy. This has raised an important question that whether quantum AI/ML algorithms are also as vulnerable as classical AI/ML models, giving birth to a new field of research known as quantum adversarial AI/ML.  This book will provide a comprehensive overview of the research in the emerging field of quantum adversarial AI/ML and include seminal work from world-leading quantum and classical AI/ML experts. The book aims to cover an up-to-date summary of the cutting-edge recent work on advancing quantum AI/ML and its benchmarking against adversarial attacks. This will be the first such book on the topic and will provide an excellent reference for graduate students and industry experts who are interested in quantum AI/ML for security sensitive autonomous systems.  The book will also cover quantum robustness beyond AI/ML to provide a comprehensive understanding of the field of research.