Selective prediction-set models with coverage guarantees
This work addresses uncertainty quantification for complex tasks like medical prediction, offering a method to ensure coverage guarantees, though it appears incremental as it builds on existing ideas from decision theory and robust likelihood.
The paper tackles the problem of unreliable uncertainty quantification in black-box predictors by proposing selective prediction-set models that output prediction sets or abstain when uncertain, and it shows improved performance on ICU patient data with accurate coverage inference.
Though black-box predictors are state-of-the-art for many complex tasks, they often fail to properly quantify predictive uncertainty and may provide inappropriate predictions for unfamiliar data. Instead, we can learn more reliable models by letting them either output a prediction set or abstain when the uncertainty is high. We propose training these selective prediction-set models using an uncertainty-aware loss minimization framework, which unifies ideas from decision theory and robust maximum likelihood. Moreover, since black-box methods are not guaranteed to output well-calibrated prediction sets, we show how to calculate point estimates and confidence intervals for the true coverage of any selective prediction-set model, as well as a uniform mixture of K set models obtained from K-fold sample-splitting. When applied to predicting in-hospital mortality and length-of-stay for ICU patients, our model outperforms existing approaches on both in-sample and out-of-sample age groups, and our recalibration method provides accurate inference for prediction set coverage.