Back to researchers
Nicholas Carlini
Adversarial ML, security of deployed models
High-signal work on real failure modes: adversarial examples, extraction, and practical model security.
Highlights
SecuritySafetyRed teaming
Focus: Adversarial ML, security of deployed models
Why it matters: High-signal work on real failure modes: adversarial examples, extraction, and practical model security.
Start here
Research Areas
SecuritySafetyRed teaming