Adversarial Machine Learning

Techniques that exploit vulnerabilities in machine learning models by crafting inputs designed to cause the model to make incorrect predictions.