Article

Vision-Language-Delta-Action: Using Compact Deltas as an Executable State-Transition Interface for Robotic Manipulation

VLDA proposes modeling task-relevant state transitions with compact Delta variables, replacing VLA's direct observation-to-action mapping and the full future-state prediction of world-action models. It learns Delta under a unified objective via LT-Filter and GT-Anchor, and DDAR injects Delta into the action denoising process as a structured seed. Experiments implemented in the open-source training framework LoongForge show that VLDA improves long-horizon consistency, robustness, and generalization.

Part of the event: VLDA proposes modeling robot state transitions with compact Delta

This page provides a summary and source link only.