🔢 mathematics

Radon transform for $GL_n(\mathbb{F}_q)$

This paper investigates the Radon transform associated with unipotent radical subgroups of $GL_n(\mathbb{F}_q)$ , focusing on its properties and providing a detailed description of its eigenvalues for the specific cases of $GL_2(\mathbb{F}_q)$ and $GL_3(\mathbb{F}_q)$ .

Original authors: Ivan Motorin, Kai Yamashita

Published 2026-06-17

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Ivan Motorin, Kai Yamashita

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 a giant, complex puzzle made of numbers and shapes, where the pieces are all the possible ways to arrange a grid of numbers (matrices) over a specific type of math world called a "finite field." This paper is about a specific tool the authors built to sort, mix, and analyze these puzzle pieces. They call this tool the Radon Transform.

Here is a breakdown of what they did, using everyday analogies:

1. The Setting: The Matrix City

Think of the group $GL_n(F_q)$ as a massive city where every building is a unique, invertible grid of numbers. The authors are interested in specific neighborhoods within this city called unipotent radical subgroups.

The Analogy: Imagine these neighborhoods are like "upper-right" districts where the buildings have a very specific, rigid structure (like a staircase going up).
The authors look at two different versions of these districts: one that is the "mirror image" of the other.

2. The Tool: The Radon Transform (The Great Mixer)

The Radon Transform is essentially a machine that takes a list of items from one neighborhood and shuffles them into the other neighborhood based on how they connect.

The Analogy: Imagine you have a room full of people (the first neighborhood). You ask everyone to shout out a list of everyone they know in the second room. The Radon Transform is the act of gathering all those shouts and creating a new list for the second room.
The authors then run this process in reverse (from the second room back to the first) and combine the two steps. This creates a loop: Room A $\to$ Room B $\to$ Room A.

3. The Goal: Finding the "Secret Frequencies" (Eigenvalues)

When you run this mixing machine over and over, certain patterns emerge. Some patterns stay the same, some get bigger, and some get smaller. The authors wanted to find the exact "growth rates" or eigenvalues of this machine.

The Analogy: Think of the machine as a musical instrument. When you pluck a string, it vibrates at a specific pitch. The authors wanted to know exactly what pitches (numbers) this mathematical instrument can play.
The Big Discovery: They found that for small cities ( $n=2$ and $n=3$ ), the only possible "pitches" are powers of a number $q$ (like 1, $q$ , $q^2$ , $q^3$ , etc.). It's as if the machine can only play notes that are perfectly tuned to the size of the field it lives in.

4. How They Solved It: Breaking the Puzzle

Solving this for a huge city is impossible, so they broke it down:

The "Small City" Case ( $n=2$ ): They solved the puzzle for a 2x2 grid first. They proved the machine only produces three specific sounds: 1, $q$ , and $q^2$ .
The "Medium City" Case ( $n=3$ ): They moved to a 3x3 grid. This was much harder. To solve it, they used special mathematical "dictionaries" called Hecke Algebras.
- The Analogy: Instead of trying to count every person in the city, they translated the problem into a different language (Hecke Algebra) where the rules are simpler. In this new language, the mixing machine looks like a set of building blocks. They rearranged these blocks to see what the final shape (the eigenvalues) would be.
The Result: For the 3x3 case, they found the machine can produce powers of $q$ up to $q^6$ (and potentially 0).

5. The "Mirror" Trick

The paper uses a clever trick involving symmetry. They realized that the complex problem of mixing these specific neighborhoods could be broken down into smaller, simpler problems involving just two blocks (like splitting a 3x3 grid into a 2x2 and a 1x1 piece).

The Analogy: If you want to know how a complex dance works, you can sometimes just watch the two main dancers and ignore the background chorus. The authors showed that the behavior of the big machine is just a combination of the behaviors of these smaller, simpler machines.

6. The Big Guess (Conjecture)

After solving the small and medium cases, the authors made a bold guess for the future:

The Claim: They believe that no matter how big the city gets (any size $n$ ), the Radon Transform will always only produce "pitches" that are powers of $q$ (like $q^0, q^1, q^2...$ ).
They haven't proven this for giant cities yet, but the pattern holds perfectly for the ones they checked.

Summary

In short, this paper is about taking a complex mathematical mixing machine, breaking it down into manageable pieces using special algebraic dictionaries, and discovering that it only produces very specific, clean numbers (powers of $q$ ). The authors have solved the puzzle for small and medium-sized versions and are confident the pattern holds for all sizes, though the proof for the largest versions is still a work in progress.

Technical Summary: Radon Transform for $GL_n(\mathbb{F}_q)$

Problem Statement
This paper investigates the Radon transform associated with the unipotent radical subgroups of the general linear group $GL_n(\mathbb{F}_q)$ over a finite field $\mathbb{F}_q$ . Specifically, the authors study the operator $\mathcal{R} = R' \circ R$ , which acts on the space of complex-valued functions on the cosets $GL_n(\mathbb{F}_q)/U_\lambda$ , where $U_\lambda$ is the unipotent radical of a parabolic subgroup corresponding to a partition $\lambda$ of $n$ . The primary objective is to determine the eigenvalues of this operator. While the eigenvalue problem is known for the standard Borel subgroup (where eigenvalues are powers of $q$ ), the behavior for general partitions $\lambda$ remains less understood.

Methodology
The authors employ a combination of representation theory, Hecke algebra theory, and explicit combinatorial calculations:

Hecke Algebra Reduction: The Radon transform is identified as an endomorphism of the subspace of right $U_\lambda$ -invariants in the group algebra $\mathbb{C}[GL_n(\mathbb{F}_q)]$ . The authors utilize the structure of Yokonuma-Hecke algebras and Gelfand-Graev Hecke algebras to decompose the problem.
Decomposition via Idempotents: For specific cases, the identity element is decomposed using orthogonal idempotents associated with characters of the unipotent subgroup. This allows the Radon transform to be expressed as a sum of operators acting on smaller, commuting subalgebras.
Explicit Calculation for Small $n$ : The paper focuses on the cases $n=2$ and $n=3$ . For $n=2$ , the minimal polynomial of the operator is derived explicitly. For $n=3$ , the authors restrict the operator to specific Hecke subalgebras (Yokonuma and Gelfand-Graev types) and utilize computer algebra systems (Wolfram Mathematica) to verify minimal polynomials and characteristic equations.
Combinatorial Coefficient Analysis: For the parabolic case $\lambda = (k, n-k)$ , the authors derive explicit formulas for the matrix coefficients of the Radon transform by analyzing double cosets and counting the number of solutions to specific matrix equations over $\mathbb{F}_q$ .

Key Contributions and Results

General Properties:
- The authors prove that the Radon transform $\mathcal{R}$ has real eigenvalues, and the multiplicity of every eigenvalue is divisible by $q-1$ .
- They establish that the eigenvalue problem for a general partition $\lambda$ can be reduced to the case of partitions of the form $(k, n-k)$ .
- It is shown that the eigenvalues are non-zero algebraic integers dividing powers of $q$ .
Case $GL_2(\mathbb{F}_q)$ :
- The paper proves that the only possible eigenvalues for the Radon transform in this case are $1, q,$ and $q^2$ .
- Explicit multiplicities for these eigenvalues are calculated based on the dimension of the space of right $U(1,1)$ -invariant elements.
Case $GL_3(\mathbb{F}_q)$ :
- For the partition $\lambda = (2,1)$ , the eigenvalues of the Radon transform restricted to the Yokonuma-Hecke algebra are shown to be a subset of $\{1, q, q^2, q^3, q^4\}$ .
- For the Gelfand-Graev Hecke algebra (associated with non-trivial characters), the eigenvalues are a subset of $\{1, q, q^2, q^3\}$ .
- Combining these results, the possible eigenvalues for the full Radon transform on $GL_3(\mathbb{F}_q)$ are identified as a subset of $\{1, q, q^2, q^3, q^4, q^5, q^6\}$ (with 0 potentially appearing in specific invariant subspaces).
Parabolic Case $\lambda = (k, n-k)$ :
- The authors provide a basis description for the Hecke algebra $\mathcal{H}(GL_n(\mathbb{F}_q), U(k, n-k))$ in terms of linear maps between subspaces.
- They derive a formula for the matrix coefficients of the Radon transform acting on this algebra, involving products of $q$ -binomial-like terms.

Significance and Claims
The paper claims to provide the first explicit description of the eigenvalues of the Radon transform for $GL_n(\mathbb{F}_q)$ beyond the standard Borel case, specifically for $n=2$ and $n=3$ . The authors demonstrate that in these cases, the eigenvalues are powers of $q$ .

Based on their findings, the authors formulate Conjecture 7.6, stating that for any partition $\lambda$ , the eigenvalues of the Radon transform acting on $\mathbb{C}[GL_n(\mathbb{F}_q)/U_\lambda]$ are powers of $q$ . The paper positions the case $\lambda = (k, n-k)$ as central to proving this conjecture for general $n$ , suggesting that understanding the Yokonuma-Hecke algebra structure is key to resolving the general problem. The work relies on the structural properties of unipotent Hecke algebras and does not propose new applications outside the scope of representation theory and the spectral analysis of these specific operators.