Roots of real-valued zero mean maps: Compositions of linear functionals and equivariant maps
Francesca Cantor and Julia D’Amico and Florian Frick and Eric Myzelev · 2025
Read Roots of real-valued zero mean maps: Compositions of linear functionals and equivariant maps by Francesca Cantor and Julia D’Amico and Florian Frick and Eric Myzelev (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2503-19729.txt.
Introduction
Where a smooth function crosses zero can be boxed in by symmetry, and a new framework tightens those boxes. Francesca Cantor and co-authors build that framework in this 2025 paper. The functions in focus average to zero and come from pairing a fixed symmetric map with simple linear readings. Trigonometric sums are a leading case.
The framework bounds how the zero set can spread and how rich its shape can be. Those bounds widen and, in places, sharpen known results about where such sums keep one sign. A wider family of function spaces fits the same frame as special cases. The team also uses the frame to prove that efficient averaging rules exist for these paired functions. Such rules matter for accurate numerical integration.
Catalog record
| CREATOR | Francesca Cantor and Julia D’Amico and Florian Frick and Eric Myzelev |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 58 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2503-19729
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Roots of real-valued zero mean maps: Compositions of linear functionals and equivariant maps Francesca Cantor and Julia D’Amico and Florian Frick and Eric Myzelev Abstract . We develop a novel topological framework that yields results constraining the distribution of zeros of certain zero mean real-valued maps, namely those obtained from composing a fixed equivariant map with linear functionals. We use this framework to establish upper bounds for the topology of set systems in the domain where (multivariate) trigonometric polynomials do not change their sign, generalizing and, in certain regimes, strengthening results in the literature. Our results more generally contain restrictions on the distribution of zeros of Chebyshev spaces as special cases. Lastly, we apply this framework to derive existence results for efficient cubature rules for compositions of affine functionals and equivariant maps. 1. Introduction For a finite set {S\subset\mathbb{Z}_{>0}} , a trigonometric polynomial with spectrum S is a function of the form f(t)=\sum_{k\in S}(a_{k}\sin(2\pi kt)+b_{k}\cos(2\pi kt)), where for each k\in S , a_{k} and b_{k} are real numbers that are not both zero. The degree of f is defined as the maximum of S . Since every such function is periodic, we may consider them as functions from the circle S^{1}=[0,1]/(0\sim 1) of circumference one to \mathbb{R} . Moreover, any such function has zero mean , that is, the integral over its domain vanishes. Babenko [ 6 ] showed that any trigonometric polynomial of degree at most d has a zero on any closed interval of length \frac{d}{d+1} . Our goal here is to develop a topological framework for such “distribution of zeros” problems and extend Babenko’s result in various ways to a more general class of zero mean maps. Crucial to Babenko’s approach is that the vector space T_{d} spanned by the functions \sin(2\pi\cdot kt),\cos(2\pi\cdot kt),\ \text{for}\ k\in\{1,2,\dots,d\},\ \text{and constant functions} forms a Chebyshev space , that is, a (2d+1) -dimensional vector space of continuous functions S^{1}\to\mathbb{R} containing all constant functions such that for all non-constant f\in V and b\in\mathbb{R} the preimage f^{-1}(b) has size at most d . In fact, Babenko proves more general results for Chebyshev spaces, which are important objects of study in approximation theory; see for instance [ 27 ] . A curve \gamma in \mathbb{R}^{d} is called convex if every affine hyperplane intersects its image in at most d points. It is easily seen that T_{d} is a Chebyshev space is equivalent to the convexity of the trigonometric moment curve \gamma_{d}(t)=(\sin 2\pi t,\cos 2\pi t,\sin 2\pi\cdot 2t,\cos 2\pi\cdot 2t,\dots,\sin 2\pi\cdot dt,\cos 2\pi\cdot dt). The trigonometric moment curve \gamma_{d} is also a prototypical example of a symmetric curve in addition to being convex. Indeed, it is an orbit of a circle action on Euclidean space that is free away from the origin. Such symmetry is strictly more general than the convexity of the curve (see Lemma 4.1 ), and here we show that Babenko’s result extends to the symmetric setting; see Theorem 1.1 . In Theorem 1.2 we record results for multivariate trigonometric polynomials, which once again extends to the symmetric setting; see Theorem 5.2 . Similarly, we extend a result of Adams, Bush, and the third author on sign changes of raked trigonometric polynomials on small diameter sets to this symmetric setting; see Theorem 1.3 . Sign changes of systems of functions are closely related to efficient numerical integration; we state a result about the existence of cubature rules for multivariate trigonometric polynomials and more general functions derived from the symmetric case; see Theorem 1.4 . We now carefully state these results. Trigonometric polynomials of degree d are compositions of linear functionals with \gamma_{d} . We generalize results to the setting of compositions of linear functionals with equivariant (i.e., symmetry-preserving) maps. For a group G acting on metric spaces X and Y , a map f\colon X\to Y is G -equivariant if it commutes with the G -actions, that is, if f(g\cdot x)=g\cdot f(x) for all x\in X and for all g\in G . The p -element cyclic group \mathbb{Z}/p acts on the circle S^{1} by rotations. A point x\in X is a fixed point of a G -action on X if g\cdot x=x for all g\in G . For G -actions on \mathbb{R}^{d} , we do not necessarily need that G acts by linear maps, only that the action preserves rays through the origin. Theorem 1.1 . Let \gamma\colon S^{1}\to\mathbb{R}^{2d} be a \mathbb{Z}/p -equivariant map for some prime p\geq 2 and some \mathbb{Z}/p -action on \mathbb{R}^{2d} whose only fixed point is 0 . Then there is a closed interval I\subset S^{1} of length \frac{d}{d+1} such that all compositions of \gamma with linear functionals, that is, maps f_{a}\colon S^{1}\to\mathbb{R} given by f_{a}(x)=\langle a,\gamma(x)\rangle for a\in\mathbb{R}^{2d} , have a zero on I . In fact, the origin is a convex combination of at most d+1 points in the image of \gamma . The bound of \frac{d}{d+1} for the length of intervals in Theorem 1.1 is optimal since Babenko’s original result is. Theorem 1.1 guarantees the existence of one interval I of length \frac{d}{d+1} , where every f_{a} has a zero, while Babenko’s theorem for trigonometric polynomials asserts this for every such interval; the statement for all intervals follows from Theorem 1.1 since by the circle symmetry of \gamma_{d} all closed intervals of a given length are the same up to symmetry; see Remark 3.2 . We derive structural results for the zeros of compositions of linear functionals with (\mathbb{Z}/p)^{n} -equivariant maps from the n -torus; see Theorem 5.2 . Here we first state these results in the special case of multivariate trigonometric polynomials , that is, linear combinations of the functions (t_{1},\dots,t_{n})\mapsto\cos(2\pi\sum\alpha_{i}t_{i}) and (t_{1},\dots,t_{n})\mapsto\sin(2\pi\sum\alpha_{i}t_{i}) for non-negative integers \alpha_{i} that are not all zero. The set S_{f} of all n -tuples of coefficients (\alpha_{1},\dots,\alpha_{n}) is the spectrum of the multivariate trigonometric polynomial f . Multivariate trigonometric polynomials are zero mean maps (S^{1})^{n}\to\mathbb{R} . Earlier results of this form are due to Kozma and Oravecz [ 21 ] and Steinerberger [ 28 ] ; see Subsection 2.1 . In the following the torus (S^{1})^{n} carries the \ell_{2} -metric inherited from the usual Euclidean metric on [0,1]^{n} . We prove (see Section 5 ): Theorem 1.2 . Let n\geq 1 be an integer, and let p\geq 3 be a prime. Let S\subset(\mathbb{Z}_{\geq 0})^{n}\setminus\{0\} be a set of cardinality at most \frac{1}{2}p^{n}-1 , and let r=\sqrt{\frac{1}{4}(n-1)+(\frac{p-2}{2p})^{2}} . Further assume that for every \alpha\in S some coordinate \alpha_{i} is not divisible by p . Then every multivariate trigonometric polynomial with spectrum S has a zero on any closed geodesic ball of radius r in (S^{1})^{n} . Such results do not trivially follow from the 1 -dimensional case by restricting the inputs to (t_{1},0,\dots,0) : The restriction of a multivariate trigonometric polynomial to a single non-zero variable is in general not a trigonometric polynomial, but differs from a trigonometric polynomial by an additive constant; in particular, this restriction generally does not have zero mean. We say that a function f\colon X\to\mathbb{R} changes sign on A\subset X if there are x,y\in A with f(x)\geq 0 and f(y)\leq 0 . We prove the following variant of Theorem 1.1 (see Section 4 ): Theorem 1.3 . Let \gamma\colon S^{1}\to\mathbb{R}^{2d} be a \mathbb{Z}/p -equivariant map