English

Convergence bounds for empirical nonlinear least-squares

Numerical Analysis 2021-05-13 v5 Numerical Analysis Probability

Abstract

We consider best approximation problems in a nonlinear subset M\mathcal{M} of a Banach space of functions (V,)(\mathcal{V},\|\bullet\|). The norm is assumed to be a generalization of the L2L^2-norm for which only a weighted Monte Carlo estimate n\|\bullet\|_n can be computed. The objective is to obtain an approximation vMv\in\mathcal{M} of an unknown function uVu \in \mathcal{V} by minimizing the empirical norm uvn\|u-v\|_n. We consider this problem for general nonlinear subsets and establish error bounds for the empirical best approximation error. Our results are based on a restricted isometry property (RIP) which holds in probability and is independent of the nonlinear least squares setting. Several model classes are examined where analytical statements can be made about the RIP and the results are compared to existing sample complexity bounds from the literature. We find that for well-studied model classes our general bound is weaker but exhibits many of the same properties as these specialized bounds. Notably, we demonstrate the advantage of an optimal sampling density (as known for linear spaces) for sets of functions with sparse representations.

Keywords

Cite

@article{arxiv.2001.00639,
  title  = {Convergence bounds for empirical nonlinear least-squares},
  author = {Martin Eigel and Reinhold Schneider and Philipp Trunschke},
  journal= {arXiv preprint arXiv:2001.00639},
  year   = {2021}
}

Comments

32 pages, 18 figures; major revisions

R2 v1 2026-06-23T13:01:50.070Z