Results for "state-action value"
Observing model inputs/outputs, latency, cost, and quality over time to catch regressions and drift.
Maintaining two environments for instant rollback.
Agent reasoning about future outcomes.
Agents communicate via shared state.
Control using real-time sensor feedback.
Control without feedback after execution begins.
Mathematical framework for controlling dynamic systems.
The physical system being controlled.
System returns to equilibrium after disturbance.
Finding control policies minimizing cumulative cost.
High-fidelity virtual model of a physical system.
Learning physical parameters from data.
Modeling environment evolution in latent space.
Hard constraints preventing unsafe actions.
Closed loop linking sensing and acting.
Mathematical guarantees of system behavior.
Modeling chemical systems computationally.
No agent benefits from unilateral deviation.
No agent can improve without hurting another.
Sudden jump to superintelligence.