Results for "evaluation protocol"
Deceptive Alignment
Advanced
Model behaves well during training but not deployment.
Distribution Shift
Intermediate
Train/test environment mismatch.
Model Documentation
Intermediate
Required descriptions of model behavior and limits.
Change Management
Intermediate
Governance of model changes.
Medical Imaging AI
Intermediate
AI applied to X-rays, CT, MRI, ultrasound, pathology slides.
Sensitivity
Intermediate
Ability to correctly detect disease.
False Negative
Intermediate
Failure to detect present disease.
FDA Clearance
Intermediate
US approval process for medical AI devices.