Zero-Shot Learning is an AI approach that enables models to learn to recognise new classes or concepts without explicit training examples. This is achieved by leveraging existing knowledge and transferring it to unseen classes. Zero-Shot Learning is useful when acquiring labelled data for all possible classes is challenging or unfeasible.
Adversarial machine learning involves studying and defending AI models against attacks or adversarial examples designed to deceive the system. By understanding vulnerabilities and deploying…