Ariadne: trajectory generation from state-only data, no trajectory training (NeurIPS 2026)
du_yilun · x · 2026-10-19
@SilongYong introduces Ariadne (NeurIPS 2026): exploring whether generative models can plan trajectories without ever training on trajectory data — generating trajectories from state-only data. The approach could bypass costly trajectory collection for planning and robotics tasks.
More from Embodied
- TALOS robots for solar plant maintenance show Europe's applied AI path — ingliguori · 2026-10-19
- Musk predicts an 'age of Amazing Abundance' where anyone can have anything they imagine — XFreeze · 2026-10-19
- Robot leg actuators finally working, more details teased soon — _Stocko_ · 2026-10-19
- Tesla shareholder letter touts Cybercab production, Optimus lines, 10M+ vehicles delivered — XFreeze · 2026-10-19
- DHH's $6 Shenzhen ride in a Chinese EV with S-Class interior sparks alarm for EU automakers — LinusEkenstam · 2026-10-19
- Inspire Robots launches RH37D1, a direct-drive three-finger hand that lifts up to 43kg — yogthos · 2026-10-19