Free Information Center
The projected gradient descent (PGD) attack is a widely used adversarial method in machine learning that iteratively perturbs input data to fool neural networks while ensuring perturbations remain within a specified constraint. It is a cornerstone technique for testing model robustness against adversarial examples.
Homomorphic encryption for AI refers to the application of cryptographic techniques that enable computations on encrypted data, supporting privacy-preserving artificial intelligence models. This approach allows AI systems to process sensitive information without exposing the underlying data, addressing critical security and privacy concerns.
Adversarial machine learning is a field focused on understanding and mitigating vulnerabilities in machine learning models caused by maliciously crafted inputs designed to deceive them. It studies how adversaries can manipulate data to cause errors in prediction or classification, and develops defenses to improve robustness.