Real-Time EXPO-FT: async VLA proposals plus RL critics unlock real-time π0.5
philfung · x · 2026-09-22
- Real-Time EXPO-FT makes slow VLAs usable in real time by splitting inference into two paths:
- Async: a frozen VLA generates candidate proposal chunks
- Sync: two small RL-tuned nets (an edit policy and a Q-critic) adjust and select proposals using current observations
- This unlocks π0.5 on dynamic tasks such as balancing a ball on a plate and striking a ball into a goal.
- A practical route to deploying heavyweight VLA policies on real-time robots.
Related event: Stanford's Real-Time EXPO-FT Makes VLA Robots Real-Time(2 posts)→
More from Embodied
- UBTECH sells 400 U1 Pro humanoids to Yongda Auto in ~$10M deal — CyberRobooo · 2026-11-12
- Baidu World 2026 to showcase Apollo Go robotaxi rides and Apollo World physical AI model — Baidu_Inc · 2026-11-12
- NVIDIA's open-source Alpamayo 2 Edge VLA driving model lands in Autoware — drmapavone · 2026-11-12
- A closer look at Astra's robot gripper design (and a joke at its expense) — _Stocko_ · 2026-11-12
- Robot startup passes first presale milestone with 10+ units claimed — R4nsack · 2026-11-12
- Runway launches Praxis-1, a 27B MoE world action model, with a 73ms Fast variant — ChongZzZhang · 2026-11-12