Tsinghua's EMERGE-Policy dispatches role-based sub-agents to combine VLA, world models and verifiers
jiqizhixin · x · 2026-09-21
Tsinghua University SIGS (Prof. Li Xiu's team) presents EMERGE-Policy, a multi-agent asynchronous concurrent framework that breaks the "single policy decides everything" paradigm in robotics.
Instead of forcing one VLA model to handle every step, EMERGE-Policy lets a main agent dispatch role-based sub-agents that call VLA, WAM, world models, and verifiers from a unified skill library of standardized tool interfaces — each model does what it does best, with world models predicting consequences and verifiers checking safety.
More from Embodied
- Stanford RL method fixes VLA latency, lifting robot success from 42% to 97% with 10 minutes of data — burny_tech · 2026-09-21
- Deadlifting 160kg at LUDUS: a striking strength demo — AIFlow_ML · 2026-09-21
- DeformSmith generates physics-grounded deformable assets for robot manipulation — Can Li · 2026-09-21
- RewardAI launches OM-1 robot foundation model trained on human data alone, zero-shot across arms and humanoids — zipengfu · 2026-09-21
- First human vs humanoid robot T800 fight opens a new entertainment genre — CyberRobooo · 2026-09-21
- Meta to launch camera-less smart glasses at Connect as privacy backlash grows — Scobleizer · 2026-09-21