Metraipolis › archive › dc_367c33723fec

Signed Lexical Confidence for Risk-Calibrated Intent Routing — arXiv:2610.00262

Source
arXiv (en)
Original
https://arxiv.org/abs/2610.00262
License (as archived)
CC0-1.0
Archived
2026-10-06T12:08:15.750Z
SHA-256 of the archived text
8a991ece67a5395c857ed486c267737c1f4851db43845733f21191fc80904dd1
Stable URL of this page
https://metraipolis.com/documento/dc_367c33723fec

How to cite

Signed Lexical Confidence for Risk-Calibrated Intent Routing — arXiv:2610.00262 — arXiv (en), https://arxiv.org/abs/2610.00262 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_367c33723fec (Metraipolis archive dc_367c33723fec, sha256 8a991ece67a5395c857ed486c267737c1f4851db43845733f21191fc80904dd1)
BibTeX
@misc{dc367c33723fec,
  title = {Signed Lexical Confidence for Risk-Calibrated Intent Routing — arXiv:2610.00262},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2610.00262},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_367c33723fec; sha256 8a991ece67a5395c857ed486c267737c1f4851db43845733f21191fc80904dd1},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.00262. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Selective intent routing allows an assistant to act on reliable predictions while deferring uncertain requests. Standard confidence scores primarily reflect the base model's representation, leaving an opportunity to incorporate complementary evidence without changing its decisions. We introduce a signed lexical gate that combines a sentence classifier's logit margin with a sparse lexical model's support for the classifier's predicted intent. By assigning positive evidence to lexical agreement and negative evidence to a lexically favored competing intent, the gate retains more information than either unsigned lexical confidence or a hard agreement rule. An independent binomial calibration stage selects an operating threshold for a specified risk target. Across ten runs on BANKING77, CLINC150, and HWU64, the proposed score reduces area under the risk-coverage curve by 15.8%, 15.1%, and 11.8% relative to a learned semantic-only gate. At a nominal 5% error target, it increases accepted coverage by 1.83 and 5.14 percentage points on BANKING77 and HWU64, while CLINC150 is already near full coverage. At a stricter 2% target, the simultaneous binomial procedure yields a nonempty policy in all 30 dataset-run combinations at the available calibration budgets. Matched controls show that the proposed feature improves average error ranking over the tested unsigned lexical-confidence feature, with dataset-dependent gains over binary agreement. The resulting two-feature gate provides a compact, interpretable confidence enhancement for risk-calibrated intent routing while preserving the base classifier's predictions.

Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-10-02 in cs.AI).
Categorie arXiv: Computation and Language (cs.CL) ; Artificial Intelligence (cs.AI).
Pagina del lavoro: https://arxiv.org/abs/2610.00262 — 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 2026-10-06T10:21:28.032Z dalla pagina https://arxiv.org/catchup/cs.AI/2026-10-02?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).