Two strongly convex polynomial penalties satisfy the real-line model
ProvedPGDRevision20261113.polynomial_penalty_modelsbregman-geometryconvex-optimizationformalization
The two penalty-curvature profiles on are
For either choice, use , penalty curvature , and generator curvature , with and inverse stepsize . The error is and its norm is the supremum on the entire real line.
For every , the ordinary polynomials
have second derivatives and . Both curvature profiles are at least one, and . In particular, with constant loss curvature one, both data sets satisfy all real-line classification hypotheses, including differentiability, the strict derivative signs, and growth at infinity.
Related results. Curvature-relation reduction · Geometry classification (Theorem 5.1)
Preamble
import Definitions.Def_PGDRevision20261113_Examples open PGDRevision20261113
Formal statement
theorem PGDRevision20261113.polynomial_penalty_models : PolynomialPenaltyModelsClaim := by sorry
Source
Author manuscript, Exact Residual Bounds and Optimal Geometry Selection for Proximal Gradient Methods, revised reading edition, 13 November 2026; Section 6.2, Real-line penalties and curvature relations; sections/06_classification_applications.tex, sec:real_examples.