Fei-Fei Li team unveils FACT, a force-aware foundation model for robot dexterity
stanfordnlp · x · 2026-10-21
Stanford researchers (Li Fei-Fei, Yejin Choi, Jiajun Wu, et al.) released FACT (arXiv:2610.19155), a multimodal force-action-track foundation model for human-to-robot dexterous manipulation.
- Motivation: human demos scale cheaply but lack force measurements, which dexterous manipulation critically needs.
- Introduces EgoForce, a large-scale human tactile dataset with force-rich manipulation experience, co-trained with robot demos.
- Architecture: a Track-Guided Coarse Planner (jointly predicting actions, future video, and visual tracks for cross-embodiment transfer) plus a Force-Action Joint Refiner that predicts future force and refines coarse actions.
- Evaluated on 10 tasks needing fine-grained force feedback — tool use, musical instruments, precision lab procedures — working with as few as 10 robot trajectories plus 90 human trajectories.
More from Embodied
- NVIDIA open-sources MotionBricks: 350K motion skills at 15,000 FPS, 2ms latency — aftahi_ai · 2026-10-21
- FLoTS-SLAM uses optical-flow residuals to hit 0.73 DM-IoU in dynamic monocular SLAM — zhenjun_zhao · 2026-10-21
- Plato ring launches at $199 to turn near-silent speech into AI commands — TinfoilTricorn · 2026-10-21
- Dream-1 controlled 31 unseen robot types across 53 countries within 24 hours — gharik · 2026-10-21
- Stanford's T2-z lets robots probe hidden physics with a poke, lifting success from 34% to 73% — stanfordnlp · 2026-10-21
- Snap Hires Robert Downey Jr. as Chief Customer Experience Officer to Sell Specs AR Glasses Against Meta — adariostrange · 2026-10-21