Data poisoning
Definition
A type of attack in which malicious or manipulated data is introduced into a training dataset to corrupt the behavior of a machine learning model, causing it to produce incorrect, biased, or harmful outputs.
443 plain-English definitions, from 3G to zero-day exploits. Each term explains how it works, why it matters, and where ZeroTier fits.
A type of attack in which malicious or manipulated data is introduced into a training dataset to corrupt the behavior of a machine learning model, causing it to produce incorrect, biased, or harmful outputs.