Reproduce “Convex Regression at Scale: Geometry-Preserving Constraint Sparsification” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Anish Senapati, Jose Blanchet
The authors replace convex regression’s quadratic constraint set with K nearest-observation constraints per supporting plane. For fixed dimension d>4, they prove this localized fit preserves the full estimator’s population-risk rate for general and strongly convex truths. In a synthetic d=5 case, its risk ratio falls from 696 to 1.06 as K rises from 19 to 348; this is not runtime evidence.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.