Model Security

Inversion Attack

Attack that approximately reconstructs sensitive training data by analyzing the model's outputs, threatening the confidentiality of information used for its learning.

← 뒤로