GK Question

technology hard fill_blank

The technique that adds calibrated noise to model outputs to prevent membership inference is called ________ Privacy.

Answer: Differential

Differential Privacy mathematically guarantees that model outputs don't reveal whether specific individuals were in training data. Critical for privacy-preserving ML deployment.

Topic AI/ML Security
Exam Relevance Banking, UPSC, SSC