Metraipolis › archive › dc_208dd044dc4f
Evidence-Aligned Local Composition of Discrete Experts for Sequence Restoration — arXiv:2609.05801
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2609.05801
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-16T03:15:58.274Z
- SHA-256 of the archived text
333e653ab3c90475006e7e50567a3ea97f219e37f739c33cafb68fdd288916a3- Stable URL of this page
https://metraipolis.com/documento/dc_208dd044dc4f
How to cite
Evidence-Aligned Local Composition of Discrete Experts for Sequence Restoration — arXiv:2609.05801 — arXiv (en), https://arxiv.org/abs/2609.05801 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_208dd044dc4f (Metraipolis archive dc_208dd044dc4f, sha256 333e653ab3c90475006e7e50567a3ea97f219e37f739c33cafb68fdd288916a3)
BibTeX
@misc{dc208dd044dc4f,
title = {Evidence-Aligned Local Composition of Discrete Experts for Sequence Restoration — arXiv:2609.05801},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2609.05801},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_208dd044dc4f; sha256 333e653ab3c90475006e7e50567a3ea97f219e37f739c33cafb68fdd288916a3},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.05801. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
A document modeled as a discrete sequence of tokens can be thought of as being generated from a composition of texts from different domains; a README file, for example, moves between prose, code, and configuration. When such a document is corrupted and only frozen domain experts are available, restoring it requires deciding both what is missing and which expert to trust at each position, at test time and without region labels or a trained router. We introduce evidence-aligned local composition, which infers a soft, position-wise weighting over the experts from the marginal evidence of the corrupted observation under a given corruption model, estimating the evidence from the experts' own denoising losses and smoothing the weights across positions. Because the weighting is soft, it recovers a mixture when the true composition is mixed and concentrates on one expert when that suffices. Across a categorical simulator, byte-level experts, and experts fine-tuned from a $1.3$B discrete flow-matching model, the inferred weights track the true regions at $0.85$ field accuracy on naturally mixed scientific documents, and at $0.98$ on constructed mixtures whose regions are lexically disjoint. Restoration improves over a single global weight when the experts are genuinely distinct and reduces to it when they converge, tracking a measure of expert separation. Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-09-14 in cs.LG). Categorie arXiv: Artificial Intelligence (cs.AI) ; Machine Learning (cs.LG). Pagina del lavoro: https://arxiv.org/abs/2609.05801 — 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 2026-10-16T03:05:44.088Z dalla pagina https://arxiv.org/catchup/cs.LG/2026-09-14?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).