Metraipolis › archive › dc_ad4e35bedf82

Return or Revise? Learning When Revision Helps Retrieval-Augmented QA — arXiv:2609.30087

Source: arXiv (en)
Original: https://arxiv.org/abs/2609.30087
License (as archived): CC0-1.0
Archived: 2026-10-08T04:49:59.374Z
SHA-256 of the archived text: cd0af0aafa572ca0abf065d01d377a7f36c7041de01b774134c932edf2047cfc
Stable URL of this page: https://metraipolis.com/documento/dc_ad4e35bedf82

How to cite

Return or Revise? Learning When Revision Helps Retrieval-Augmented QA — arXiv:2609.30087 — arXiv (en), https://arxiv.org/abs/2609.30087 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_ad4e35bedf82 (Metraipolis archive dc_ad4e35bedf82, sha256 cd0af0aafa572ca0abf065d01d377a7f36c7041de01b774134c932edf2047cfc)

BibTeX

@misc{dcad4e35bedf82,
  title = {Return or Revise? Learning When Revision Helps Retrieval-Augmented QA — arXiv:2609.30087},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2609.30087},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_ad4e35bedf82; sha256 cd0af0aafa572ca0abf065d01d377a7f36c7041de01b774134c932edf2047cfc},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.30087. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

We consider the decision of whether to return an existing draft answer or revise it using retrieved evidence, as in answer-revision systems. Draft confidence estimates whether the current answer is correct, but the decision requires estimating the effect of a specified revision. For offline training and evaluation, we grade both the returned draft and its candidate revision under the same correctness judge, which makes repair, harm, and the gap to an oracle observable. We call this paired effect its recoverability, and we train policies to predict it before revision. On 25,870 held-out open-domain questions across three revision setups, a scorer trained on the paired outcome has greater area under the accuracy--revision-rate curve than a matched draft-correctness scorer in all nine Llama setup--seed fits, and gains 0.23--0.68 accuracy points on average at development-selected thresholds, a difference significant across training runs only for dense retrieval. The resulting policy improves on always revising and on average closes more than a third of the oracle gap, although it still applies 38--46% of the harmful revisions. When a draft-free standard-RAG answer is also available, however, choosing between the draft and that answer is stronger by about two points for Llama and four for OLMo, and adding candidate revision as a third option yields no significant gain. Recoverability describes one revision; its value as an available action also depends on the alternatives.

Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-09-25 in cs.CL).
Categorie arXiv: Computation and Language (cs.CL) ; Information Retrieval (cs.IR); Machine Learning (cs.LG).
Pagina del lavoro: https://arxiv.org/abs/2609.30087 — 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 2026-10-08T04:27:20.126Z dalla pagina https://arxiv.org/catchup/cs.CL/2026-09-25?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).