GaP Uses Gemini Flash-Lite to Synthesize Robot Policy Graphs; G1 Learns Safety-Thinking
DynamicWebPaige · x · 2026-11-10
Google's DynamicWebPaige highlights two robotics papers: GaP (Graph-as-Policy, Spotlight S6) uses Gemini 3.1 Flash-Lite as both coding agent and VLM backbone to synthesize and self-refine executable computation graphs for robot tasks, combining ROS2/GraspGen skill nodes with Isaac Sim and Newton sim self-learning across 8 benchmark tasks (4 sim + 4 real). A second poster post-trains Gemini Robotics-ER 1.5 (SFT + RL) for safety-thinking so a Unitree G1 spots physical hazards and intervenes in real time.
More from Embodied
- Newton: An open-source GPU-accelerated physics engine built on NVIDIA Warp — ChongZzZhang · 2026-11-11
- From quantum to AI: Ge Yang on CoRL's boom and where robot learning goes next — EpisodeYang · 2026-11-11
- Newton-FBA ships sim2real cloth manipulation, learned in minutes — ChongZzZhang · 2026-11-11
- Perceptron open-sources EgoHome, a sub-second egocentric household-video benchmark — lukas_m_ziegler · 2026-11-11
- Anthropic touts Fable 5.2 robot control gains on RoboDojo; real-world box folding disappoints — _Stocko_ · 2026-11-11
- AI2 open-sources ToyBlocks to benchmark robotic world models by task-relevant fidelity — allen_ai · 2026-11-11