Preprint
Aug 2026
Adaptive Regularization for Random Features: A Neighboring Early-Stopping Rule with Oracle-Rate Guarantees
This work proposes a neighboring early-stopping rule for adaptive regularization in KRR with random features (KRR-RF), using a grid that is uniform in inverse regularization and compares only adjacent estimators, reducing the number of discrepancy comparisons relative to standard all-pairs Lepskii-type procedures.
Caihong Wang, Zhi-Bo Chen, Yue Wang
· 0 citations