Sharp estimates for the covering numbers of the Weierstrass fractal kernel
Douglas Azevedo, Karina Gonzalez, Thais Jord\~ao

TL;DR
This paper introduces a Weierstrass fractal kernel, characterizes its associated RKHS, and provides sharp estimates for the covering numbers of its unit ball, revealing the space's dense subset of nowhere differentiable functions.
Contribution
It offers the first explicit characterization of the RKHS for a Weierstrass fractal kernel and derives sharp covering number estimates for its unit ball.
Findings
RKHS contains a dense subset of continuous nowhere differentiable functions
Sharp estimates for covering numbers of the RKHS unit ball are provided
The space's properties highlight its fractal and complex structure
Abstract
In this paper, we use the infamous continuous and nowhere differentiable Weierstrass function as a prototype to define a Weierstrass fractal kernel. We investigate the properties of the reproducing kernel Hilbert space (RKHS) associated with this kernel by presenting an explicit characterization of this space. In particular, we show that this space has a dense subset composed of continuous but nowhere differentiable functions. Moreover, we present sharp estimates for the covering numbers of the unit ball of this space as a subset of the continuous functions.
Peer Reviews
No public reviews on file for this paper yet. If you reviewed it on a platform where reviews are public (OpenReview, ICLR, NeurIPS, ICML), you can paste yours below so the community can read it here.
Videos
No videos yet. Explain this paper in a talk, walkthrough, or lecture? Add one.
Taxonomy
TopicsMathematical Dynamics and Fractals · Mathematical Approximation and Integration · Advanced Numerical Analysis Techniques
