Metraipolis › archive › dc_072a599a7367

Word-Level Text Unmixing via Evidence-Preserving Ownership Routing with Language Models — arXiv:2610.06603

Source
arXiv
Original
https://arxiv.org/abs/2610.06603
License (as archived)
CC0-1.0
Archived
2026-10-06T05:27:47.477Z
SHA-256 of the archived text
29f9ee7ce328b2874b0e31218f9f70fe108d1d655f4c0e968dfc6c84f0b8214c
Stable URL of this page
https://metraipolis.com/documento/dc_072a599a7367

How to cite

Word-Level Text Unmixing via Evidence-Preserving Ownership Routing with Language Models — arXiv:2610.06603 — arXiv, https://arxiv.org/abs/2610.06603 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_072a599a7367 (Metraipolis archive dc_072a599a7367, sha256 29f9ee7ce328b2874b0e31218f9f70fe108d1d655f4c0e968dfc6c84f0b8214c)
BibTeX
@misc{dc072a599a7367,
  title = {Word-Level Text Unmixing via Evidence-Preserving Ownership Routing with Language Models — arXiv:2610.06603},
  howpublished = {arXiv},
  url = {https://arxiv.org/abs/2610.06603},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_072a599a7367; sha256 29f9ee7ce328b2874b0e31218f9f70fe108d1d655f4c0e968dfc6c84f0b8214c},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv, https://arxiv.org/abs/2610.06603. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Text from multiple sources can become interleaved into a single sequence when attribution metadata is lost, such as overlapping speech transcripts, document reading flows, or concurrent agent streams. We formalize this challenge as Word-Level Text Unmixing: given an interleaved lexical stream and source count K, recover the original source sequences while preserving every word occurrence and its within-source order exactly. Directly generating separated texts with LLMs can omit, duplicate, or hallucinate words, violating this exact-reconstruction objective. We therefore propose Evidence-Preserving Ownership Routing (EPOR), which decouples source-ownership prediction from reconstruction. EPOR adapts a causal LLM to predict canonical ownership routes conditioned on the mixed stream and prior routing decisions. At inference, completion-safe constrained decoding is combined with deterministic indexed reconstruction, yielding structurally valid K-source partitions that preserve every observed occurrence exactly once. We also introduce UNMIXBENCH, covering controlled synthetic mixtures, timestamp-derived speech from AMI and ICSI, layout-derived document streams from ReadingBank, and simulated concurrent digital outputs. Across five evaluation tracks, a 4B EPOR model achieves the lowest mean minimum-permutation word error rate among finetuned baselines, reducing the five-track mean by 22.3% relative to compact source-array generation and remaining competitive with zero-shot frontier LLMs. These results show that when lexical evidence is fully observed, separating ownership inference from lexical regeneration provides a reliable alternative to direct generation.

Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.AI, nuovi lavori del giorno).
Categorie arXiv: Computation and Language (cs.CL) ; Artificial Intelligence (cs.AI).
Pagina del lavoro: https://arxiv.org/abs/2610.06603 — 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 2026-10-06T05:24:06.177Z dalla pagina https://arxiv.org/list/cs.AI/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).