From the course: Security Risks in AI and Machine Learning: Categorizing Attacks and Failure Modes
Join today to access over 24,800 courses taught by industry experts.
ML adversarial training
From the course: Security Risks in AI and Machine Learning: Categorizing Attacks and Failure Modes
ML adversarial training
- [Narrator] Living in outer space is not easy. There's no gravity and no option to step out for a walk in the fresh air. Without proper preparation, astronauts bodies would not survive the attack that being in space poses to their health. So they spend hours adjusting to being weightless in NASA's KC-135 and learn how to repair equipment while underwater to simulate the experience of being outside of the ship. The challenges of space are unusual to the human body, which is why training for them is critical to the safety of the astronauts and their mission. Adversarial training in AI and ML works a little like getting a person ready for space. The negative impacts or unusual challenges are defined often through threat modeling. Then the system is trained for resilience using defensive techniques. Models that are trained in controlled environments may not function properly in noisy real world situations and models can…
Practice while you learn with exercise files
Download the files the instructor uses to teach the course. Follow along and learn by watching, listening and practicing.
Download courses and learn on the go
Watch courses on your mobile device without an internet connection. Download courses using your iOS or Android LinkedIn Learning app.