Using Nano Banana image generation as visual goal synthesis for deformable robot manipulation
DynamicWebPaige · x · 2026-11-10
Two robot-learning papers highlighted at the conference: (1) PGRD synthesizes a goal image with Nano Banana from a scene plus instruction (e.g. "pass rope through red slot"), then executes deformable-object manipulation against that visual target, with demoed sequences for threading, lifting, and rotating tasks; (2) Being-H0.5 scales robot learning with 16,000+ hours of egocentric human video, using Gemini video understanding to produce dual-timescale labels (1s fine-grained motions + 10s intent summaries) for cross-embodiment VLAs.
More from Embodied
- Amazon Ring unveils $250 autonomous home security drone that patrols and streams live video — Polymarket · 2026-11-10
- NVIDIA-led study: unblocking VLM gradients in VLA training lifts manipulation progress from 10.3% to 32.9% — stepjamUK · 2026-11-10
- Google DeepMind launches Gemini Robotics 2.5 with 96% accuracy sorting Mercedes-Benz parts — GoogleDeepMind · 2026-11-10
- SlashVibe is a physical desk controller for managing multiple running AI coding agents — Scobleizer · 2026-11-10
- Path Robotics raises $300M Series E at $2.45B valuation to expand physical AI — Rewkang · 2026-11-10
- Unitree Go2 robot asks clarifying questions via Gemini and Bayesian active learning — DynamicWebPaige · 2026-11-10