MRP fits label-wise monotone functions that map calibrated confidence to correctness reliability, improving post-calibration reranking and fallback utility across six relevance datasets without changing the calibrated probabilities.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.IR 1years
2026 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Post-Calibration Reliability Reranking of Relevance Decisions via Label-wise Monotone Projection
MRP fits label-wise monotone functions that map calibrated confidence to correctness reliability, improving post-calibration reranking and fallback utility across six relevance datasets without changing the calibrated probabilities.