Abstract:Covering numbers are a powerful tool used in the development of approximation algorithms, randomized dimension reduction methods, smoothed complexity analysis, and others. In this paper we prove upper bounds on the covering number of numerous sets in Euclidean space, namely real algebraic varieties, images of polynomial maps and semialgebraic sets in terms of the number of variables and degrees of the polynomials involved. The bounds remarkably improve the best known general bound by Yomdin-Comte, and our proof is much more straightforward. In particular, our result gives new bounds on the volume of the tubular neighborhood of the image of a polynomial map and a semialgebraic set, where results for varieties by Lotz and Basu-Lerario are not directly applicable. We illustrate the power of the result on three computational applications. Firstly, we derive a near-optimal bound on the covering number of low rank CP tensors, quantifying their approximation properties and filling in an important missing piece of theory for tensor dimension reduction and reconstruction. Secondly, we prove a bound on the required dimension for the randomized sketching of polynomial optimization problems, which controls how much computation can be saved through randomization without sacrificing solution quality. Finally, we deduce generalization error bounds for deep neural networks with rational or ReLU activation functions, improving or matching the best known results in the machine learning literature while helping to quantify the impact of architecture choice on generalization error.

Covering dimension of finite topological spaces

Topological Characterizations for Three Covering Approximation Operators

Set multi-covering via inclusion-exclusion

Covering Dimension of Finite Distributive Lattices

The covering type of closed surfaces and minimal triangulations

Topological minimal sets and their applications

Dimension Estimation of Image Manifolds by Minimal Cover Approximation

On covering by translates of a set

No Subgroup Theorem for the Covering Dimension of Topological Groups

Dimension, depth and zero-divisors of the algebra of basic $k$-covers of a graph

Covering Number of Real Algebraic Varieties and Beyond: Improved Bounds and Applications

Minimal covers in the Weihrauch degrees

Covering Generalized Rough Sets Based on the Intersection of Minimal Description

On Sparse Covers of Minor Free Graphs, Low Dimensional Metric Embeddings, and other applications

Topological Characteristic Factors and Nilsystems

Covering Numbers in Linear Algebra

Almost covers of finite sets of points

A Non-Metrizable Compact Linearly Ordered Topological Space, Every Subspace Of Which Has A Sigma-Minimal Base

Minimal Finite Models

The Shape Of Compact Covers

Covering $B_X$ by finitely many convex sets