🔢 mathematics

Spectral Criteria for Uniqueness Pairs of Unitary Transforms

This paper proposes a spectral approach to identifying uniqueness pairs for unitary transforms by mapping two-sided sampling problems to lower-bound spectral problems of confined Hamiltonians, thereby generalizing existing Wirtinger-Poincaré-based criteria to a broader class of transforms including the fractional Fourier and Hankel transforms.

Original authors: Oleg Szehr

Published 2026-07-10

📖 6 min read🧠 Deep dive

CC BY 4.0

Original authors: Oleg Szehr

Original paper licensed under CC BY 4.0 (http://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

Imagine you are trying to solve a mystery: you have a secret message (a signal), but you can't see the whole thing at once. You only get to peek at a few specific spots in the message itself, and a few specific spots in its "shadow" (its transformed version, like a frequency map). The big question is: How many peeks do you need to be absolutely sure you've found the only possible message that fits?

For a long time, mathematicians have known the answer for the most common type of shadow (the Fourier transform), but they used a very specific, rigid rulebook to prove it. In this paper, Oleg Szehr proposes a brand-new way to look at the problem. Instead of using that old rulebook, he suggests we treat the problem like a game of quantum confinement.

The Quantum Cage Game

Here is the core idea: Imagine your secret message is a tiny, energetic particle bouncing around inside a room.

The Zeros are Walls: Every time the paper says the message is zero at a certain point, imagine a wall suddenly appears there. If the message is zero at point A and point B, the particle is trapped in a tiny cage between A and B. It can't escape.
The Shadow is a Second Room: The message also has a "shadow" (its transform). If the shadow is zero at certain points, that means the particle is also trapped in a second, invisible room made of walls in the shadow-world.

The paper argues that if you build too many walls in both the real room and the shadow room at the same time, the particle gets squeezed so tight that it simply cannot exist. It has to vanish. If the particle vanishes, the only possible message was the "empty" message (zero everywhere).

This is a new kind of Uncertainty Principle. Usually, we say you can't know a particle's position and speed perfectly at the same time. Szehr says: "If you try to trap a particle in two different ways simultaneously with too much force, the particle just disappears."

The Old Way vs. The New Way

Previously, mathematicians used a tool called the Wirtinger-Poincaré inequality to prove when the message is unique. You can think of this as a generic "energy meter" that works well for simple, flat rooms.

Szehr's paper shows that this old tool is actually just a special case of a deeper, more powerful principle.

What it rules out: The paper argues that we shouldn't just rely on the old "flat room" math for every situation. If the room has weird shapes, weights, or corners (like in advanced signal processing), the old tool might give the wrong answer.
The New Tool: Instead of a generic meter, Szehr uses Hamiltonians (the energy equations from quantum mechanics). He calculates the "ground state energy"—the lowest possible energy a particle can have in a specific cage.
- If the cage is too small or the walls are too close, the minimum energy required to keep the particle alive becomes higher than the energy the particle is allowed to have.
- Result: The particle must vanish. Therefore, the message is unique.

Testing the Theory on Different "Rooms"

The paper proves this works for the standard Fourier transform (the usual shadow), but it also stretches the idea to two other tricky scenarios:

The Fractional Fourier Transform (The Rotating Room):
Imagine the shadow isn't just a static map, but a map that has been rotated by an angle $\alpha$ .
- The paper shows that the "critical threshold" (the point where the message vanishes) changes depending on how much you rotate the room.
- If you rotate it by 90 degrees (the standard case), the rule is strict. If you rotate it less, the walls can be spaced slightly differently before the message vanishes. The math shows the critical spacing scales with $|\sin(\alpha)|$ .
The Hankel Transform (The Radial Room):
This is for signals that look like ripples in a pond (radial symmetry).
- Here, the "room" has a weird singularity at the center (like a black hole in the middle of the room).
- The paper proves that even with this weird center, if you place walls (zeros) too close together, the particle still vanishes. However, the math is different because the "floor" of the room isn't flat; it curves due to the singularity. The paper provides specific formulas for how close the walls can be in this curved space.

How Sure Are We?

The paper doesn't just guess; it proves these results.

It uses rigorous variational arguments (a method of finding the lowest possible energy in a system) to show that if the sampling points (the walls) are "uniformly supercritical" (meaning they are dense enough everywhere), the only solution is zero.
It explicitly states that for the standard Fourier transform, this new method recovers the exact same known results as the old method, but explains why they work through energy levels instead of just inequalities.
For the new transforms (Fractional and Hankel), it provides new, proven criteria for uniqueness. It doesn't just suggest them; it derives them from the spectral properties of the operators involved.

The Takeaway

The paper's main finding is that uniqueness in signal recovery is really a physics problem about confinement.

If you sample a signal and its transform at points that are too dense (creating too many "Dirichlet walls"), you force the signal's energy to exceed what is physically possible for a non-zero signal. The signal collapses to zero.

This approach is powerful because it replaces a one-size-fits-all math trick with a flexible energy-based framework. It tells us that for any unitary transform (any "shadow" system), we just need to calculate the lowest energy floor of the confined system to know exactly how many samples we need to guarantee a unique recovery. It turns a dry sampling problem into a vivid story of a particle trapped in a cage that gets too small to hold it.

Technical Summary: Spectral Criteria for Uniqueness Pairs of Unitary Transforms

Problem Statement
The paper addresses the fundamental problem of identifying "uniqueness pairs" for unitary transforms. A uniqueness pair consists of two discrete sets, $\Lambda$ (in the physical domain) and $M$ (in the transformed domain), such that prescribing the values of a function $f$ on $\Lambda$ and its transform $Uf$ on $M$ uniquely determines $f$ (typically forcing $f=0$ if all samples are zero). While recent work by Kulikov, Nazarov, and Sodin established recovery criteria for the Fourier transform using Wirtinger-Poincaré functional inequalities, this paper seeks to generalize these results to a broader class of unitary transforms (including the fractional Fourier and Hankel transforms) using a distinct spectral approach.

Methodology
The author proposes a spectral framework motivated by quantum mechanics. The core methodology interprets the zeros of a function $f$ and its transform $Uf$ as imposing Dirichlet boundary conditions in two conjugate representations. This converts the two-sided sampling problem into a problem of spectral lower bounds for "confined Hamiltonians."

The approach relies on the following structural assumptions for a unitary operator $U$ on $L^2(S)$ :

Invariant Form Domain: There exists a Hilbert space $X \subset L^2(S)$ that is invariant under $U$ ($UX=X$) and serves as the form domain of a strictly positive self-adjoint operator $\mathcal{X}$ .
Hamiltonian Structure: The operator $\mathcal{X}$ is defined via a closed quadratic form involving multiplication by the square of the variable in both the original and transformed domains: $\mathcal{X} = \frac{1}{2}(M_{x^2} + U^* M_{x^2} U)$ .
Local Compatibility: The kinetic part of the form must be compatible with local Dirichlet confinement. Specifically, if a function vanishes at the endpoints of a set of disjoint intervals, the global form energy must be bounded below by the sum of the energies of the function restricted to those intervals with Dirichlet boundary conditions.

The proof strategy involves decomposing the domain into intervals defined by the sampling sets $\Lambda$ and $M$ . By utilizing the variational characterization of the lowest eigenvalue (spectral floor) of the confined operators on these intervals, the author derives a contradiction if the sampling sets are "supercritical" (i.e., sufficiently dense) and the function is non-zero.

Key Contributions and Results

Spectral Reformulation of Fourier Uniqueness: For the standard Fourier transform, the paper recovers the known supercritical uniqueness criterion (where the product of interval length and distance from the origin is less than $\pi$ ). However, it derives this via the spectral floor of the confined harmonic oscillator and the Dirichlet Laplacian, rather than directly invoking the Wirtinger-Poincaré inequality. This reveals that the Wirtinger constant is simply the spectral floor of the confined Laplacian in this specific case.
Generalization to Unitary Transforms: The paper formulates a general theorem (Theorem 3.3) providing sufficient conditions for uniqueness pairs for any unitary transform satisfying the structural assumptions. The condition requires that the "kinetic Dirichlet confinement" be uniformly supercritical, meaning the ratio of the square root of the local spectral floor to the maximum distance of the interval endpoints from the origin exceeds 1.
Fractional Fourier Transform (FrFT): The framework is applied to the FrFT, which corresponds to a rotation in phase space. The resulting uniqueness criterion shows that the critical density threshold is rescaled by the factor $|\sin(\alpha)|$ , where $\alpha$ is the rotation angle. The paper demonstrates that as $\alpha \to 0$ or $\pi$ , the required sampling density becomes infinitely high (supercriticality requires the intervals to shrink to zero relative to their position).
Hankel Transform: The theory is extended to the Hankel transform, which arises in problems with radial symmetry. Here, the associated Hamiltonian is a radial harmonic oscillator with an inverse-square Bessel potential. The uniqueness criteria depend on the spectral floors of this specific Sturm-Liouville operator, incorporating both the Dirichlet confinement energy and the singular behavior at the origin (Friedrichs realization).

Significance and Claims
The paper claims to provide a "confinement form of the uncertainty principle": a non-zero quantum state cannot be excessively confined in two conjugate representations simultaneously. The significance of this work lies in:

Unification: It abstracts the mechanism behind uniqueness pairs, showing that the Wirtinger-Poincaré inequality is a special case of a broader principle where the critical threshold is determined by the local spectral floor of the Hamiltonian associated with the specific transform.
Physical Interpretation: It offers a physical interpretation of sampling densities, linking them to the energy levels of confined quantum systems.
Generalization: It successfully extends uniqueness criteria to transforms with non-trivial coefficients, weights, and singular endpoints (such as the Hankel transform), where standard Wirtinger-Poincaré arguments may not directly apply or require modification.

The author emphasizes that this spectral viewpoint does not merely rephrase existing results but identifies the local mechanism (the first confined eigenmode) governing the critical threshold, suggesting that the sharpness of these bounds originates from configurations where interval-wise ground-state bounds are approached asymptotically.

The Quantum Cage Game

The Old Way vs. The New Way

Testing the Theory on Different "Rooms"

How Sure Are We?

The Takeaway

More like this