Topic: AI Security and Privacy

With the widespread use of artificial intelligence (AI) technologies in various applications like autonomous driving, industrial IoT, health care, and language processing, the security and privacy issues in developing and utilizing AI systems are becoming increasingly critical. For instance, to overcome data silos caused by privacy concerns, privacy-preserving AI protocols need to be developed for multiple parties to collaborate to train a better model without leaking their private data; more robust AI models need to be built to make correct predictions against random or adversarial perturbations. Other than the privacy and security of AI systems, using AI methods to identify security vulnerabilities and help with security enhancement is also becoming an effective and important approach in designing secure systems.

The aim of this Special Issue is to bring together research investigations in identifying security and privacy issues in AI systems, building secure and robust AI algorithms and protocols, and developing AI-based security enhancement techniques. Prospective authors are invited to submit original research contributions on topics including but not limited to:
● Data, model, and membership inference attacks;
● Data and model poisoning attacks;
● Differential privacy;
● Adversarial training and certified robustness;
● Privacy and security issues in federated learning;
● Privacy-preserving data mining;
● AI applications for security and privacy;

Extended conference papers are also welcome, provided they have been revised and contain at least 50% new content (and the previously published proceedings paper is submitted with the research paper).

Artificial intelligence (AI), security, privacy, attacks, federated learning, AI applications

30 Jun 2024