arXiv · 2611.22096
COSI: A Framework for Robustly Valid Inference
Abstract
We introduce COSI (COnfidence from Scale Invariance), a framework for constructing confidence sets for statistical functionals. COSI builds on the HulC (Hull-based Confidence), which uses median unbiasedness to construct confidence sets from batch estimates. COSI uses scale-invariance properties shared by many common estimators. We divide the data into batches, compute an estimate in each batch, and retain each candidate parameter value for which the batch estimates pass a test of scale invariance. The resulting confidence sets are asymptotically valid under weaker assumptions than classical Wald and bootstrap procedures require, and their coverage can be closer to the nominal level in small samples. Applications include robust inference for the mean, regression coefficients in misspecified linear models, tensor norms, and Fréchet means.
Explore related subjects
Keep this discovery
Explore connections, maps & timelines
Arun K. Kuchibhotla, Sivaraman Balakrishnan, Larry Wasserman. 2026-11-17. COSI: A Framework for Robustly Valid Inference. https://arxiv.science/abs/2611.22096
Citations retain the original authors and DOI and include this site’s record URL. If this discovery service contributed to your research, you can also cite arXiv Science by Indiciae: BibTeX · RIS.