Hello, you are using an old browser that's unsafe and no longer supported. Please consider updating your browser to a newer version, or downloading a modern browser.
Training Camp • Cybersecurity Glossary
Data poisoning corrupts machine learning models by injecting malicious or mislabeled samples into training data to cause errors or hidden backdoors.
Data Poisoning Definition: Data poisoning corrupts machine learning models by injecting malicious or mislabeled samples into training data to cause errors or hidden backdoors.
Data poisoning is an adversarial machine learning attack in which an attacker deliberately injects, alters, or mislabels samples in a model's training data to corrupt the resulting model's behavior. Poisoning can degrade overall accuracy, create targeted misclassifications, or implant backdoors that trigger malicious outputs on attacker-chosen inputs. It is a key risk in supply chains that rely on crowdsourced, scraped, or third-party datasets.
Data Poisoning is one of the topics you'll master in the AIGP Boot Camp.
AIGP Boot Camp →