2026
Romberg-Extrapolated Zeroth-Order Gradient Estimator: Higher-Order Bias Reduction with Preserved Leading Directional Variance
ICML 2026poster
Zeroth-order optimization is widely used when gradients are unavailable, but the standard two-point estimator suffers from $\mathcal{O}(r^2)$ truncation bias at smoothing radius $r$. Existing bias-reduction schemes typically increase the leading directional variance under a fixed number of function …