Facet-0 Robotic Foundation Model
Facet-0: A Robotic Foundation Model for Contact-Rich Precise Manipulation
Facet-0 enables precise robotic assembly at sub-millimeter tolerances, outperforming existing baselines by over fivefold in success rates.
Facet-0 is a robotic foundation model that predicts and values contact consequences of actions.
It unifies multimodal representation learning and reinforcement learning using a joint action-wrench proposal.
Trained on ManuFacet-1K, a 1,000-hour corpus, it achieves 82% mean success on five sub-millimeter computer-assembly tasks, compared with 15% for the strongest baseline.
The system maintains 0.5 mm placement accuracy and 50 ms command latency.
Facet-0: A Robotic Foundation Model for Contact-Rich Precise Manipulation
Real-world robotic assembly at sub-millimeter tolerances demands spatial precision, compliant interaction, and robustness to contact failures. We present Facet-0, a robotic foundation model that predicts and values the contact consequences of its actions. Facet-0 unifies multimodal representation learning and reinforcement learning (RL) post-training around a joint action-wrench proposal: a causal wrench history is aligned with vision-language semantics and kinematic state, and flow matching generates each action chunk together with the future wrist-wrench profile it is expected to induce. Deployment rollouts train a distributional Action-Wrench Critic to distinguish motions with similar task progress but different contact outcomes, while phase-aware rewards and contact-selective credit concentrate policy improvement on decisive interactions. To accommodate part-specific dynamics, a lightweight bounded actor reuses the frozen representation for on-robot adaptation; RL remains defined over executable Cartesian actions, while an auxiliary wrench head preserves predictive, non-commanded action-contact coupling. Trained on ManuFacet-1K, a 1,000-hour force-synchronized corpus spanning three embodiments and multiple manufacturing cells, the bounded task-adapted system reaches 82% mean success on five sub-millimeter computer-assembly tasks, compared with 15% for the strongest baseline, with 0.5 mm placement accuracy and 50 ms command latency.
This report is categorized under AITOP's "Models" desk. Key entities and technical topics under continuous tracking include Facet-0, ManuFacet-1K, robotic foundation model, reinforcement learning.
Source attribution: curated from arXiv cs.LG (@Haoyuan Deng, Haichao Liu, Wenkai Guo, Yuan Ling, Zaijia Yang, Yuanjiang Xue, Haosheng Sun, Liangzi Wang, Ziwei Wang).
