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

- Source: HuggingFace Daily Papers
- Published: 2026-10-19T00:00:00.000Z
- RoboHot: https://robohot.news/en/items/yr7j7m60n7mklrehgatbj2swb
- Original: https://huggingface.co/papers/2610.25874

## Summary

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.