Metraipolis › archive › dc_8bb506265f63

Soft Symbol Grounding for Prototypical Concepts — arXiv:2609.12247

Source
arXiv (en)
Original
https://arxiv.org/abs/2609.12247
License (as archived)
CC0-1.0
Archived
2026-10-16T03:02:36.746Z
SHA-256 of the archived text
068e44f84855d179a4b564f195edf0495410208d0c3d28247ee107eb8bd3c782
Stable URL of this page
https://metraipolis.com/documento/dc_8bb506265f63

How to cite

Soft Symbol Grounding for Prototypical Concepts — arXiv:2609.12247 — arXiv (en), https://arxiv.org/abs/2609.12247 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_8bb506265f63 (Metraipolis archive dc_8bb506265f63, sha256 068e44f84855d179a4b564f195edf0495410208d0c3d28247ee107eb8bd3c782)
BibTeX
@misc{dc8bb506265f63,
  title = {Soft Symbol Grounding for Prototypical Concepts — arXiv:2609.12247},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2609.12247},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_8bb506265f63; sha256 068e44f84855d179a4b564f195edf0495410208d0c3d28247ee107eb8bd3c782},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.12247. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Neuro-symbolic models are usually trained with supervision only on final labels, leaving the intermediate concepts unobserved. Since many concept assignments are consistent with a given label, training can predict labels correctly while recovering the wrong concepts, a failure known as a reasoning shortcut. Prototypical networks reduce shortcuts by anchoring each concept to a few labeled examples, but existing methods still couple perception and reasoning through a hand-crafted, task-specific differentiable loss that must be redesigned for every task. We introduce \textbf{Soft-PNet}, which removes this loss: it reframes concept grounding as a Metropolis walk over a precomputed cache of feasible symbolic solutions, guided by a prototype distribution built from a single labeled anchor per concept, and trains against one KL objective between the prototype-weighted cache and the network's concept predictions. The objective is identical across tasks and remains applicable when the solution space cannot be enumerated. On \texttt{MNIST-EvenOdd}, Visual Sudoku, and \texttt{Kand-Logic} under scarce supervision, Soft-PNet matches loss-engineered prototypical networks at the concept and label levels and recovers concepts that soft-grounding baselines miss, with no loss engineering and lower training time.

Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-09-14 in cs.AI).
Categorie arXiv: Artificial Intelligence (cs.AI).
Pagina del lavoro: https://arxiv.org/abs/2609.12247 — 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 2026-10-16T02:51:27.765Z dalla pagina https://arxiv.org/catchup/cs.AI/2026-09-14?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).