SOURCE-LINKED INTELLIGENCE

Think Like a World Model, Act Like a VLA: Distilling World-Model Representations into Compact Robot Policies

arXiv · AI, language, vision and robotics · article · Sep 21, 2026 · UTC

Vision-Language-Action (VLA) models map observations to actions with no objective that accounts for how the world responds, so their robustness is bounded primarily by data coverage. World models carry precisely that missing objective and are better grounded for it, yet rolling the future forward costs seconds per decision and rules them out of the control loop. We show the two can be separated. What a world model knows about physical scenes lives in its internal features; generating the future is merely the objective that produced them, so the grounding can be inherited while the generative m

Read original source ↗ Open in workspace

recordType

paper

region

Global