dreamerv3

This paper diagnoses long-horizon failures in world models, attributing them to kinematic rather than dynamic imagination. The authors introduce a metric (iKCE) and show that imagined rollouts fail to capture dynamic regime changes even as policy rewards collapse.

GPLD introduces a gradient-penalized latent dynamics regularizer for DreamerV3 to enforce local smoothness in transition learning, improving sample efficiency on continuous control tasks, especially complex locomotion.