WorldGuide: Learning Success-Failure Boundaries in Latent World Models for Vision-Language-Action Policies

Latent world models offer a promising way to improve Vision-Language-Action policies by capturing the consequences of actions. However, models trained primarily on expert demonstrations have limited exposure to failure outcomes and may struggle to distinguish visually similar successful and failed i