Sem-ECE is an asymptotically unbiased calibration error estimator for open-ended QA that uses semantic sampling of answers to derive confidence from class frequencies, with two variants that diverge on hard questions.
Kagel and Dan Levin.Common Value Auctions and the Winner’s Curse
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
citation-role summary
background 1
citation-polarity summary
fields
cs.CL 1years
2026 1verdicts
UNVERDICTED 1roles
background 1polarities
background 1representative citing papers
citing papers explorer
-
A Semantic-Sampling Framework for Evaluating Calibration in Open-Ended Question Answering
Sem-ECE is an asymptotically unbiased calibration error estimator for open-ended QA that uses semantic sampling of answers to derive confidence from class frequencies, with two variants that diverge on hard questions.