Model Inversion Glossary Definition

Model inversion is an attack technique that attempts to reconstruct sensitive training data or attributes from a model's outputs or behavior.

Privacy and AI security teams evaluate model inversion risk when models are trained on sensitive data or exposed through public interfaces.

Additional Resource: Authoritative source.

Related concepts include Privacy and Model Extraction.