Metraipolis › archive › dc_ae38fb160857

Latent RL Priors for Data-Efficient Dexterous Manipulation — arXiv:2611.08458

Source
arXiv (en)
Original
https://arxiv.org/abs/2611.08458
License (as archived)
CC0-1.0
Archived
2026-11-06T14:54:14.193Z
SHA-256 of the archived text
84709a48e62bb667645fe6cb1f16a1991fe61cc169a1215ef02239eda8f614f6
Stable URL of this page
https://metraipolis.com/documento/dc_ae38fb160857

How to cite

Latent RL Priors for Data-Efficient Dexterous Manipulation — arXiv:2611.08458 — arXiv (en), https://arxiv.org/abs/2611.08458 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_ae38fb160857 (Metraipolis archive dc_ae38fb160857, sha256 84709a48e62bb667645fe6cb1f16a1991fe61cc169a1215ef02239eda8f614f6)
BibTeX
@misc{dcae38fb160857,
  title = {Latent RL Priors for Data-Efficient Dexterous Manipulation — arXiv:2611.08458},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2611.08458},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_ae38fb160857; sha256 84709a48e62bb667645fe6cb1f16a1991fe61cc169a1215ef02239eda8f614f6},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2611.08458. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Learning dexterous manipulation from limited visual demonstrations is challenging because policies must acquire both perceptual grounding and precise control, while reinforcement learning (RL) in simulation can acquire rich control behaviors from large-scale interaction with privileged information. We introduce \textbf{Latent RL Prior}, a two-stage framework that reuses this control knowledge for data-efficient visual policy learning. During privileged pretraining, we distill large-scale RL rollouts into a flow-matching action decoder conditioned on proprioception and latent task context. During visual adaptation, we freeze the decoder and learn a visual interface that maps images to its latent context using action supervision alone, without privileged labels or latent targets. The same decoder can further be shared across arms by learning coordinated visual contexts for bimanual manipulation. Across three simulated tasks, our method outperforms full-parameter fine-tuning of $\pi_{0.5}$ and GR00T N1.7 in the low-data regime; on door lifting and bimanual assembly, it surpasses full-data baselines using $10\times$ and $5\times$ fewer demonstrations, respectively. Further analysis shows that direct visual-to-latent conditioning improves adaptation performance and training stability over routing visual predictions through the pretrained state encoder. Robustness to unseen mass variation and greater linear accessibility of contact-force information in decoder representations further provide evidence of retained RL-acquired control knowledge. Project website: this https URL .

Tema degli esploratori: statistica e scienza dei dati (elenco arXiv cs.RO, nuovi lavori del giorno).
Categorie arXiv: Robotics (cs.RO).
Pagina del lavoro: https://arxiv.org/abs/2611.08458 — 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 2026-11-06T14:25:57.574Z dalla pagina https://arxiv.org/list/cs.RO/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).