Results for "system regulation"

Models accessible only via service APIs.

Sensitivity of a function to input perturbations.

Ensuring AI systems pursue intended human goals.

Model exploits poorly specified objectives.

Maximizing reward without fulfilling real goal.

Model optimizes objectives misaligned with human values.

Ensuring learned behavior matches intended objective.

Learned subsystem that optimizes its own objective.

Model behaves well during training but not deployment.

Maintaining alignment under new conditions.

US framework for AI risk governance.

Ability to inspect and verify AI decisions.

Dynamic resource allocation.

Storing results to reduce compute.

Devices measuring physical quantities (vision, lidar, force, IMU, etc.).

Algorithm computing control actions.

Using output to adjust future inputs.

Artificial environment for training/testing agents.

Space of all possible robot configurations.

Detecting and avoiding obstacles.

Estimating robot position within a map.

Humans assist or override autonomous behavior.

Controlling robots via language.

Closed loop linking sensing and acting.

AI applied to X-rays, CT, MRI, ultrasound, pathology slides.

Testing AI under actual clinical conditions.

Models estimating recidivism risk.

Legal right to fair treatment.

AI proposing scientific hypotheses.

AI selecting next experiments.