Stanford's FACT model uses human force signals to beat T-Rex by up to 60 points on dexterous tasks
stanfordnlp · x · 2026-10-21
Stanford researchers (Yejin Choi, Li Fei-Fei, Jiajun Wu et al.) released FACT, a multimodal force-action-track foundation model for human-to-robot dexterity.
- Cascades a track-guided coarse planner into a force-action joint refiner
- Fine-tuned on just 10 robot and 90 human trajectories
- Beats T-Rex by up to 60 points across ten dexterous tasks
- The force-signal mismatch between humans and robots is acknowledged and tested; a Pith AI review finds the empirical case for force supervision clear, backed by a new dataset
More from Embodied
- Amazon FAR's FIBER uses action-aligned latent planning to let Unitree G1 open doors and carry boxes — ChongZzZhang · 2026-10-21
- Sentdex trains robots to go from lying down to a half kneel, eyeing a full stand — Sentdex · 2026-10-21
- NVIDIA open-sources MotionBricks: 350K motion skills at 15,000 FPS, 2ms latency — aftahi_ai · 2026-10-21
- NVIDIA Cosmos3-H-Dreams: Generative Simulation Brings Clinical Surgery to Life — Hugging Face Blog · 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