Evaluating Real-World Adversarial ML Attack Risks and Effective Management: Robustness vs Non-ML Mitigations
The MLSecOps Podcast
00:00
Adversarial ML Techniques: Model Inversion and Differential Privacy
The speakers discuss the techniques of model inversion and differential privacy in adversarial ML. They explain the concept of model inversion as stealing information from a model, and discuss how differential privacy can limit information leakage. They also mention their own research in differential privacy and its application in protecting data.
Play episode from 24:09
Transcript
