technology hard Fill in the Blank

The technique that protects ML models by adding calibrated noise to outputs to prevent inference attacks is called ________ Privacy.

  1. Nature-based / Biological
  2. margin
  3. Differential
  4. Account Aggregator

Answer: Differential

Differential Privacy adds mathematical noise to model outputs, providing provable privacy guarantees against membership inference and model inversion attacks.

Topic AI/ML Security: Advanced Threats
Exam Relevance Banking, UPSC, SSC
More technology Questions Back to all questions