Security -

: Subtly altering input data to trick a model into making incorrect predictions.

The intersection of security and deep learning covers two primary areas: using deep learning to security (e.g., intrusion detection) and protecting deep learning models from vulnerabilities (e.g., adversarial attacks) . Key Security Threats to Deep Learning security

: Reverse-engineering a trained model to reveal its parameters or architecture. : Subtly altering input data to trick a