Tag
This paper addresses objective mismatch in model-based RL by proposing offline diagnostics to predict closed-loop performance of latent world models. On LunarLander-v3, the Reward Observability Fraction (ROF) and a Composite score (CROF) enable selecting checkpoints that yield strong MPC and model-based RL policies with far fewer real-environment interactions.
The failure of Blue Origin's New Glenn rocket has catastrophic consequences for NASA's Artemis program, delaying lunar lander development and increasing reliance on SpaceX's Starship.