Metraipolis › archive › dc_f78cf48a93bf

Specification-Aware Predictive Uncertainty Decomposition — arXiv:2611.02956

Source
arXiv (en)
Original
https://arxiv.org/abs/2611.02956
License (as archived)
CC0-1.0
Archived
2026-11-03T11:08:35.442Z
SHA-256 of the archived text
cfe453e940d1768bee32b4f321467d308b58d829a154cccb63703b44d54eaa6d
Stable URL of this page
https://metraipolis.com/documento/dc_f78cf48a93bf

How to cite

Specification-Aware Predictive Uncertainty Decomposition — arXiv:2611.02956 — arXiv (en), https://arxiv.org/abs/2611.02956 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_f78cf48a93bf (Metraipolis archive dc_f78cf48a93bf, sha256 cfe453e940d1768bee32b4f321467d308b58d829a154cccb63703b44d54eaa6d)
BibTeX
@misc{dcf78cf48a93bf,
  title = {Specification-Aware Predictive Uncertainty Decomposition — arXiv:2611.02956},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2611.02956},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_f78cf48a93bf; sha256 cfe453e940d1768bee32b4f321467d308b58d829a154cccb63703b44d54eaa6d},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2611.02956. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Classical decompositions of predictive uncertainty, such as the information-theoretic one, are defined for a fixed representation $x$ of the input object. They do not consider that $x$ may be incomplete, nor do they measure how much the missing information can affect the prediction. In Large Language Models, users can clarify an underspecified input without any retraining. The clarified input can describe the user's unobserved intent $z^*$ more accurately and change the predictive uncertainty about the outcome. In this paper, we formalize this setting probabilistically and derive a three-term decomposition of prediction risk into specification, residual aleatoric, and epistemic uncertainties. We show that at the original information level, when only $x$ is available, the aleatoric uncertainty, represented by a Bayes risk, has two components. The first, which we call specification uncertainty, is the expected reduction in this Bayes risk from revealing the intent. The second is the residual aleatoric uncertainty, i.e., the Bayes risk that remains once the intent is known, averaged over its possible values $z$ given $x$. The distinction is important, as specification uncertainty informs the decision to request clarification before making a prediction. We build approximations of these components and connect them to existing uncertainty measures. Our experiments across different domains and modalities show distinct roles for the three components. Specification uncertainty is the most informative for detecting underspecified inputs, while residual aleatoric uncertainty is the most informative for detecting in-distribution prediction errors.

Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.AI, nuovi lavori del giorno).
Categorie arXiv: Machine Learning (stat.ML) ; Artificial Intelligence (cs.AI); Computation and Language (cs.CL); Machine Learning (cs.LG).
Pagina del lavoro: https://arxiv.org/abs/2611.02956 — 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 2026-11-03T10:59:46.146Z dalla pagina https://arxiv.org/list/cs.AI/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).