Metraipolis › archive › dc_fa8af8c79a4e

Facet-0: A Robotic Foundation Model for Contact-Rich Precise Manipulation — arXiv:2609.01596

Source
arXiv (en)
Original
https://arxiv.org/abs/2609.01596
License (as archived)
CC0-1.0
Archived
2026-10-17T06:52:41.894Z
SHA-256 of the archived text
0a0a95ac39a3a37516d84bcd086414418df74c962004c1e4501c352c439a7ef8
Stable URL of this page
https://metraipolis.com/documento/dc_fa8af8c79a4e

How to cite

Facet-0: A Robotic Foundation Model for Contact-Rich Precise Manipulation — arXiv:2609.01596 — arXiv (en), https://arxiv.org/abs/2609.01596 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_fa8af8c79a4e (Metraipolis archive dc_fa8af8c79a4e, sha256 0a0a95ac39a3a37516d84bcd086414418df74c962004c1e4501c352c439a7ef8)
BibTeX
@misc{dcfa8af8c79a4e,
  title = {Facet-0: A Robotic Foundation Model for Contact-Rich Precise Manipulation — arXiv:2609.01596},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2609.01596},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_fa8af8c79a4e; sha256 0a0a95ac39a3a37516d84bcd086414418df74c962004c1e4501c352c439a7ef8},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.01596. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

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.

Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-09-02 in cs.LG).
Categorie arXiv: Robotics (cs.RO) ; Machine Learning (cs.LG).
Pagina del lavoro: https://arxiv.org/abs/2609.01596 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito).
Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina.
Letto il 2026-10-17T06:41:55.936Z dalla pagina https://arxiv.org/catchup/cs.LG/2026-09-02?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).