Proposes a goal-oriented lower-tail calibration framework for GPs in BO using occurrence calibration and thresholded mu-calibration, with tcGP shown to preserve dense exploration while improving calibration and performance on benchmarks.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
stat.ML 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Goal-Oriented Lower-Tail Calibration of Gaussian Processes for Bayesian Optimization
Proposes a goal-oriented lower-tail calibration framework for GPs in BO using occurrence calibration and thresholded mu-calibration, with tcGP shown to preserve dense exploration while improving calibration and performance on benchmarks.