Difficulty: Advanced

What would have happened under different conditions.

Measures joint variability between variables.

Model behaves well during training but not deployment.

Agent reasoning about future outcomes.

Diffusion model trained to remove noise step by step.

Robots learning via exploration and growth.

Generative model that learns to reverse a gradual noise process.

High-fidelity virtual model of a physical system.

Models effects of interventions (do(X=x)).

Randomizing simulation parameters to improve real-world transfer.

Motion considering forces and mass.

Predicts next state given current state and action.

Vector whose direction remains unchanged under linear transformation.

AI systems that perceive and act in the physical world through sensors and actuators.

Intelligence emerges from interaction with the physical world.

System-level behavior arising from interactions.

Competition arises without explicit design.

Risk threatening humanity’s survival.

Average value under a distribution.

External sensing of surroundings (vision, audio, lidar).

Sudden jump to superintelligence.

AI reinforcing market trends.

Sudden extreme market drop.

Exact likelihood generative models using invertible transforms.

Mathematical guarantees of system behavior.

Computing end-effector position from joint angles.

Mathematical representation of friction forces.

Two-network setup where generator fools a discriminator.

Models that learn to generate samples resembling training data.

Direction of steepest ascent of a function.