Causal isotonic calibration for heterogeneous treatment effects
This provides robust calibration for heterogeneous treatment effect estimation in causal inference, though it appears incremental as it builds on existing calibration methods.
The authors tackled the problem of calibrating predictors of heterogeneous treatment effects by proposing causal isotonic calibration and cross-calibration, achieving fast doubly-robust calibration rates under weak conditions without requiring hold-out sets.
We propose causal isotonic calibration, a novel nonparametric method for calibrating predictors of heterogeneous treatment effects. Furthermore, we introduce cross-calibration, a data-efficient variant of calibration that eliminates the need for hold-out calibration sets. Cross-calibration leverages cross-fitted predictors and generates a single calibrated predictor using all available data. Under weak conditions that do not assume monotonicity, we establish that both causal isotonic calibration and cross-calibration achieve fast doubly-robust calibration rates, as long as either the propensity score or outcome regression is estimated accurately in a suitable sense. The proposed causal isotonic calibrator can be wrapped around any black-box learning algorithm, providing robust and distribution-free calibration guarantees while preserving predictive performance.