Metraipolis › archive › dc_03089aecd313

Meta-Ctrl: Guaranteed Plan Generation by Decoupling Syntactic and Semantic Constraints — arXiv:2608.22149

Source: arXiv (en)
Original: https://arxiv.org/abs/2608.22149
License (as archived): CC0-1.0
Archived: 2026-10-18T07:42:41.916Z
SHA-256 of the archived text: df88990ee3e227969cddcc9905bc0145f1189823c406c7e909dcc63e70af7cac
Stable URL of this page: https://metraipolis.com/documento/dc_03089aecd313

How to cite

Meta-Ctrl: Guaranteed Plan Generation by Decoupling Syntactic and Semantic Constraints — arXiv:2608.22149 — arXiv (en), https://arxiv.org/abs/2608.22149 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_03089aecd313 (Metraipolis archive dc_03089aecd313, sha256 df88990ee3e227969cddcc9905bc0145f1189823c406c7e909dcc63e70af7cac)

BibTeX

@misc{dc03089aecd313,
  title = {Meta-Ctrl: Guaranteed Plan Generation by Decoupling Syntactic and Semantic Constraints — arXiv:2608.22149},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2608.22149},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_03089aecd313; sha256 df88990ee3e227969cddcc9905bc0145f1189823c406c7e909dcc63e70af7cac},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2608.22149. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

LLMs generate fluent plans for robots but routinely violate the syntactic and se8mantic constraints they must satisfy to execute, and existing remedies trade formal guarantees against plan quality: soft methods (affordance scoring, grounded decoding) give no guarantee, while symbolic planners (LLM+P) discard the LM's commonsense. We propose \textbf{Meta-Ctrl}, a constrained-decoding framework that guarantees the encoded constraints while preserving the base LM's plan quality. Meta-Ctrl introduces \emph{meta-tokens}---a compact vocabulary of grounded actions---enforcing syntax at the token level and semantics (preconditions, goals, ordering) at the action level, an exact factorization that cuts the memory of constrained decoding from over 107TB to under 2GB. With it, a small open-weight LM becomes competitive where it otherwise sits at the bottom of the leaderboard: on WAH-NL under the LoTa-Bench protocol it reaches the highest reported subgoal success rate, exceeding GPT-4's, with consistent gains across the Embodied Agent Interface. We further demonstrate it on a real tabletop robot, where every generated plan satisfies its preconditions and goals by construction. Project website: this https URL

Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-08-25 in cs.AI).
Categorie arXiv: Robotics (cs.RO) ; Artificial Intelligence (cs.AI).
Pagina del lavoro: https://arxiv.org/abs/2608.22149 — 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 2026-10-18T07:37:52.859Z dalla pagina https://arxiv.org/catchup/cs.AI/2026-08-25?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).