SimVTA uses simulator-rendered video to post-train driving world action models with RL
OpenDriveLab · hf · 2026-10-16
OpenDriveLab with Shanghai Innovation Institute, Li Auto, Tongji and Tsinghua introduce SimVTA, a simulation-grounded post-training framework for driving world action models. Since driving logs only contain future videos for recorded trajectories, alternative RL-explored trajectories lack visual supervision; SimVTA samples candidate trajectories per observation, renders future videos for high-reward candidates with a real-world simulator (e.g., World Engine), and uses this video supervision alongside RL trajectory optimization.
More from Embodied
- AutoAgent0: VLM-powered agentic runtime lifts long-tail driving success by 28.3pp — yining_hong · 2026-10-16
- CarryOn: dual-loop harness adds personalization and safe recovery to frozen VLA robots — Zixuan Wang · 2026-10-16
- FoveaVLA replaces wrist cameras with learned foveation, hitting 70.8% success — Haoyu_Xiong_ · 2026-10-16
- Musk teases two-way Grok x Optimus integration: Grok in the bot, bot in Grok — beffjezos · 2026-10-16
- Robotics data startups compared: Scale AI tops $1B revenue as 9 LLM vs robotics data suppliers benchmarked — FinanceYF5 · 2026-10-16
- Kroger cuts 3,000 corporate jobs (37%) for AI; humanoid robots next in the 2030s — csuwildcat · 2026-10-16