🔢 mathematics

Supersingular Drinfeld modules, Brandt matrices, and rank-metric codes

This paper establishes a stabilization formula for the dimension of morphism spaces between supersingular Drinfeld modules using Brandt matrices and automorphic $L$ -functions, and leverages this result to construct semifield rank-metric codes via a hyperplane-avoidance argument.

Original authors: Giacomo Micheli, Mihran Papikian

Published 2026-04-21

📖 6 min read🧠 Deep dive

CC BY 4.0

Original authors: Giacomo Micheli, Mihran Papikian

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

The Big Picture: Building a Digital Fortress

Imagine you are an architect trying to build a digital fortress (a cryptographic code) that is incredibly hard to break. In the world of mathematics, these fortresses are called Rank-Metric Codes. They are used to protect data in things like internet security, cloud storage, and secure communications.

To build the strongest fortress, you need a special kind of "bricks" (mathematical structures) that fit together perfectly. If the bricks are too loose, the fortress collapses. If they are too tight, you can't build anything at all.

This paper is about finding the perfect bricks using a very exotic type of mathematical object called a Drinfeld Module. Specifically, the authors focus on a special, "super-strong" version called a Supersingular Drinfeld Module.

1. The Building Blocks: Drinfeld Modules

Think of a Drinfeld Module as a magic machine.

You put a number into the machine.
The machine performs a very specific, twisted calculation on it (using something called a "twisted polynomial").
The output is a new number.

These machines are defined over a "finite field," which is like a universe with a limited number of stars (numbers). In this universe, you can't count to infinity; you just cycle through a fixed set of numbers.

2. The "Super" Machines: Supersingular Drinfeld Modules

Most of these magic machines are ordinary. But some are Supersingular.

Analogy: Imagine a standard car engine vs. a Ferrari engine. The Ferrari (Supersingular) has a much more complex internal structure.
Why it matters: Because they are so complex, they have a huge "engine room" (called an Endomorphism Ring) full of extra parts that can interact with the machine in unique ways. The authors discovered that these extra parts are the key to building better codes.

3. The Problem: The "Missing Brick"

The authors wanted to use these super-machines to build a perfect code.

The Goal: They needed a specific number of "bricks" (mathematical dimensions) to make a code that is mathematically perfect (called an MRD code).
The Issue: When they counted the bricks they had in their "engine room," they found they were short by a few.
- Analogy: You need 100 bricks to build a wall. You have 98. You can't build the wall perfectly; there's a gap.

4. The Discovery: The Stabilization Formula

The authors spent a lot of time counting these bricks for different sizes of machines. They found a surprising pattern:

If you look at machines of a certain size, the number of bricks you have follows a strict rule once the machine gets big enough.
The Rule: The number of bricks is always 2 × (Size + 1) - (Deficit).
The "Deficit": This deficit depends on the "character" of the universe (the prime number $p$ ).
The Breakthrough: They proved that for large enough machines, this deficit is constant. It doesn't keep growing or changing randomly. It stabilizes.

Why is this cool? It means we can predict exactly how many bricks we have, no matter how big the machine gets. We know exactly how far off we are from the perfect number.

5. The Tool: Brandt Matrices (The "Map")

To count these bricks, the authors used a tool called Brandt Matrices.

Analogy: Imagine a giant subway map or a flight schedule.
- The stations are the different "Supersingular Machines."
- The lines connecting them represent the "paths" (isogenies) you can take from one machine to another.
The Brandt Matrix is a table that tells you exactly how many paths exist between any two machines for a specific trip length.
The authors developed a new, faster algorithm to draw this map. Instead of checking every single path one by one (which is slow), they used the unique properties of the "twisted polynomials" to calculate the map instantly.

6. The Solution: Fixing the Gap

Now that they knew exactly how many bricks they were missing, they had to fix the wall. They used two different strategies depending on the "Deficit":

Strategy A (Odd Deficit): The "Zero Entry" Trick
- If the deficit is odd, they looked at their "Map" (Brandt Matrix) for a specific connection that has zero paths.
- Analogy: If you are trying to build a wall between two cities, and the map says "No roads exist between City A and City B," you can use that empty space to your advantage. By choosing a pair of machines that don't talk to each other, the "bad" bricks disappear, and the remaining bricks form a perfect, solid wall.
Strategy B (Even Deficit): The "Good Hyperplane" Trick
- If the deficit is even, you can't just find a zero path. Instead, you have to be a master sculptor.
- Analogy: Imagine you have a block of clay (your collection of bricks) that is slightly too big. Some parts of the clay are "bad" (they crack the wall). You need to slice off a thin layer (a hyperplane) that removes all the bad clay but keeps the good stuff.
- The authors proved that even though the bad clay is scattered, there is almost always a way to slice the block so that the remaining piece is perfect.

7. The Result: A New Generation of Codes

By combining these strategies, the authors successfully built Semifield Codes.

These are not just any codes; they are Maximum Rank Distance (MRD) codes.
Real-world impact: These codes are the "gold standard" for error correction. They are used to ensure that data sent over noisy networks (like deep space communication or crowded Wi-Fi) arrives perfectly intact, even if parts of the message get scrambled.

Summary

The Problem: We need perfect mathematical bricks to build secure data codes, but our "super-machines" (Drinfeld modules) usually give us slightly too few or too many.
The Discovery: The authors found a rule that predicts exactly how many bricks we have, no matter how big the machine gets.
The Tool: They built a fast "map" (Brandt Matrices) to navigate between these machines.
The Fix: They found clever ways to slice away the "bad" bricks or choose specific machine pairs to create a perfect, gap-free wall.
The Outcome: We now have a reliable method to generate the strongest possible codes for protecting digital information.

In short, they took a very abstract, high-level math problem involving "twisted polynomials" and "quaternion algebras" and turned it into a practical recipe for building unbreakable digital locks.

1. Problem Statement

The paper addresses two interconnected problems in arithmetic geometry and coding theory:

Dimension Stabilization: Determining the exact $F_q$ -dimension of the space of morphisms between supersingular Drinfeld modules of rank 2, filtered by their $\tau$ -degree. While the dimension of such spaces is known to be bounded, the precise behavior for large degrees (specifically the "stable" range) was not fully characterized for general supersingular modules.
Construction of Rank-Metric Codes: Utilizing the structure of Drinfeld modules to construct semifield codes (a specific type of Maximum Rank Distance or MRD code). The authors aim to overcome the dimension deficit that arises when constructing these codes from supersingular modules over fields where the characteristic prime has degree $d > 1$ .

2. Methodology

The authors employ a synthesis of arithmetic geometry, the theory of automorphic forms, and computational algebra:

Drinfeld Module Theory: They analyze supersingular Drinfeld modules of rank 2 over finite fields $k = \mathbb{F}_p$ (where $p$ is a prime of degree $d$ in $A = \mathbb{F}_q[T]$ ). They utilize the fact that the endomorphism ring of such a module is a maximal order in a quaternion division algebra $D$ ramified at $p$ and $\infty$ .
Brandt Matrices: They define Brandt matrices $B(m)$ which encode the arithmetic of isogenies between supersingular modules. The entry $b_{i,j}(m)$ counts the number of isogenies of norm $m$ between modules $\phi_i$ and $\phi_j$ , weighted by automorphism groups.
Theta Series and Automorphic Forms: The core theoretical tool is the decomposition of theta series associated with these modules. They relate the counting of morphisms to the Fourier coefficients of harmonic cochains on the Bruhat-Tits tree of $PGL_2(F_\infty)$ $P G L_{2} (F_{\infty})$ .
- They decompose the space of harmonic cochains into Eisenstein series and cusp forms.
- They leverage deep results by Drinfeld, Deligne, and Grothendieck regarding the $L$ -functions of cuspidal forms, specifically that the $L$ -function of a cuspidal form is a polynomial in $q^{-s}$ of degree $d-3$ .
Algorithmic Approach: They develop an efficient algorithm to compute Brandt matrices and the spaces of morphisms $M_s(\phi, \psi)$ . This relies on the twisted polynomial ring $k\{\tau\}$ , where isogenies are elements, and norm testing is reduced to right-division in this non-commutative ring.

3. Key Contributions and Results

A. Stabilization Formula (Theorem 1.2 / Theorem 3.5)

The primary theoretical result is a precise formula for the dimension of the space of morphisms $M_s(\phi, \psi) = \{u \in \text{Hom}(\phi, \psi) \mid \deg_\tau(u) \leq s\}$ for supersingular rank-2 Drinfeld modules over $\mathbb{F}_p$ (where $\deg(p)=d$ ).

For all $s \geq d-2$ :
$\dim_{\mathbb{F}_q} M_s(\phi, \psi) = 2(s + 1) - (d - 1)$

Implication: The dimension stabilizes to a linear function of $s$ with a constant deficit of $d-1$ compared to the "ideal" dimension $2(s+1)$ required for a full semifield code.
Proof Mechanism: The proof shows that for $s \geq d-1$ , the sum of Brandt matrix entries corresponding to degree $s$ vanishes due to the properties of the cusp form $L$ -function, forcing the dimension recurrence to stabilize.

B. Efficient Algorithm for Brandt Matrices (Section 3.4)

The authors present an algorithm to compute Brandt matrices $B(q)$ for primes $q$ of degree up to $s$ .

Method: It solves a system of linear equations over $\mathbb{F}_q$ derived from the commutation relation $u\phi_T = \psi_T u$ in the twisted polynomial ring.
Complexity: The complexity is roughly $O(q^{2s} s^2 m)$ , where $m$ is the extension degree. This is specific to function fields and avoids the lattice enumeration methods used in number fields (e.g., Pizer's algorithm).
Verification: The authors provide explicit computations of Brandt matrices for various primes and degrees, verifying that they commute and satisfy consistency checks (e.g., row sums equal $|q|+1$ ).

C. Construction of Semifield Codes (Section 4)

The paper applies the stabilization result to construct semifield codes (MRD codes with parameters $m=n=d_{min}$ ).

The Challenge: For $d \geq 2$ , the dimension of $M_s$ is $2(s+1) - (d-1)$ , which is strictly less than the required $2 \cdot \deg(q)$ for a code on $\phi[q]$ where $\deg(q) = s+1$ .
Strategies:
1. Odd $d$ : Choose a prime $q$ such that $2\deg(q) = 2(s+1) - (d-1)$ . If the Brandt matrix $B(q)$ has a zero entry at $(\phi, \psi)$ , then $M_s(\phi, \psi)$ contains no non-invertible elements (isogenies with kernel), yielding a valid semifield code.
2. Even $d$ : The dimension parity is wrong for the direct approach. Instead, they search for a "good hyperplane" (a codimension-1 subspace) within $M_s(\phi, \psi)$ that contains no "bad elements" (non-invertible morphisms).
Results: The authors demonstrate via Magma computations that for $d \leq 4$ , such constructions are feasible. For $d=2$ and $d=4$ , they found that the vast majority of hyperplanes are "good," successfully producing semifield codes.

4. Significance

Theoretical Unification: The paper bridges the gap between the arithmetic of Drinfeld modules (Brandt matrices, $L$ -functions) and coding theory (Rank-metric codes). It provides a rigorous explanation for the dimension behavior of morphism spaces, which was previously only observed empirically or in specific cases.
New Code Constructions: It offers a systematic method to generate semifield codes from supersingular Drinfeld modules, expanding the known families of MRD codes. This is particularly relevant for applications in network coding and cryptography.
Computational Advances: The proposed algorithm for computing Brandt matrices in function fields is more efficient and structurally distinct from classical number field algorithms, offering a new tool for researchers in arithmetic geometry over function fields.
Generalization Potential: The authors conjecture a generalization of the stabilization formula for rank $r$ modules: $\dim M_s = r(s+1) - \frac{r(r-1)(d-1)}{2}$ , suggesting a deeper structural pattern yet to be proven.

In summary, this work provides a complete theoretical framework for understanding the morphism spaces of supersingular Drinfeld modules and leverages this understanding to construct new, efficient rank-metric codes, validated by explicit computational examples.