Adversarial Attacks

Context poisoning is an adversarial technique that manipulates the context provided to AI models to produce biased outputs.

Curriculum poisoning involves manipulating training data to degrade AI model performance.

The Fast Gradient Sign Method is a technique for generating adversarial examples in machine learning.

A Leakage Attack exploits vulnerabilities in AI systems to extract sensitive information from models or training data.

Model Injection is a type of attack that manipulates AI models by injecting malicious inputs to alter their behavior.

Model poisoning is an attack that compromises machine learning models by introducing malicious data.