Metraipolis › archive › dc_fbfdaa4bb1df

A polynomial-size logical twin for neural networks — arXiv:2611.01146

Source: arXiv (en)
Original: https://arxiv.org/abs/2611.01146
License (as archived): CC0-1.0
Archived: 2026-11-03T11:45:53.959Z
SHA-256 of the archived text: e5128e32f29f415e86cebb8f5c39d5b5350eeab6312f2f4510d5725a795466dd
Stable URL of this page: https://metraipolis.com/documento/dc_fbfdaa4bb1df

How to cite

A polynomial-size logical twin for neural networks — arXiv:2611.01146 — arXiv (en), https://arxiv.org/abs/2611.01146 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_fbfdaa4bb1df (Metraipolis archive dc_fbfdaa4bb1df, sha256 e5128e32f29f415e86cebb8f5c39d5b5350eeab6312f2f4510d5725a795466dd)

BibTeX

@misc{dcfbfdaa4bb1df,
  title = {A polynomial-size logical twin for neural networks — arXiv:2611.01146},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2611.01146},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_fbfdaa4bb1df; sha256 e5128e32f29f415e86cebb8f5c39d5b5350eeab6312f2f4510d5725a795466dd},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2611.01146. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

A logical twin of a neural network is a logical representation that mimics the computation of the network through logical formulas. We revisit two earlier proposals for constructing logical twins of feedforward neural networks in Łukasiewicz logic, both of which are susceptible to exponential blow-up in running time and representation size. Building on ideas from both approaches, we present a new construction that directly yields logical twins in Łukasiewicz logic. We show that this construction can be computed in polynomial time and produces polynomial-size logical twins. Finally, we report an empirical evaluation of the three methods, showing that the new proposal substantially outperforms the other two.

Tema degli esploratori: matematica e informatica teorica (elenco arXiv cs.LO, nuovi lavori del giorno).
Categorie arXiv: Logic in Computer Science (cs.LO).
Pagina del lavoro: https://arxiv.org/abs/2611.01146 — 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 2026-11-03T11:08:07.223Z dalla pagina https://arxiv.org/list/cs.LO/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).