🔢 mathematics

Image closure of symmetric wide-matrix varieties

This paper establishes that the Zariski closure of the image of a Sym( $\mathbb{N}$ )-equivariant morphism between symmetric wide-matrix varieties is defined by finitely many Sym( $\mathbb{N}$ )-orbits and possesses the Sym( $\mathbb{N}$ )-Noetherian property, ensuring that every descending chain of Sym( $\mathbb{N}$ )-stable closed subsets stabilizes.

Original authors: Jan Draisma, Rob H. Eggermont, Azhar Farooq, Leandro Meier

Published 2026-08-12

📖 8 min read🧠 Deep dive

CC BY 4.0

Original authors: Jan Draisma, Rob H. Eggermont, Azhar Farooq, Leandro Meier

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 organize a massive, chaotic library. In this library, the books aren't just arranged by author or title; they are arranged by a magical rule where every time you add a new shelf, the entire library automatically rearranges itself to keep everything symmetrical. This is the world of "infinite-dimensional varieties" in mathematics, a branch of algebraic geometry that deals with shapes defined by equations involving an endless number of variables. Usually, when you have an infinite number of variables, things get messy and unpredictable; you can't easily list all the rules that define a shape because the list might go on forever. However, mathematicians have discovered that if you impose a specific kind of symmetry—like the way a snowflake looks the same no matter how you rotate it—you can sometimes tame this chaos. The key question is: if you take a simple, well-behaved shape and project it into this complex, infinite world, does the resulting shape stay manageable? Can we describe it with a finite list of rules, even though the space it lives in is infinite?

This paper, titled "Image Closure of Symmetric Wide-Matrix Varieties," tackles exactly that question. The authors, Jan Draisma, Rob H. Eggermont, Azhar Farooq, and Leandro Meier, prove that when you take a specific type of simple mathematical object (a matrix with a fixed number of rows but a growing number of columns) and map it into a more complex, multi-dimensional space (like a tensor, which is a generalization of a matrix to higher dimensions) while respecting symmetry, the resulting shape is surprisingly well-behaved. They show that even though the space is infinite, the "shadow" or "image" of this map is defined by only a finite number of repeating patterns of rules. Furthermore, they prove that this shape is "Noetherian," a fancy mathematical term meaning that if you start looking for smaller and smaller pieces inside it, you will eventually stop finding new ones; the process of digging deeper always comes to a halt. This is a big deal because it means these complex, symmetric shapes are not as wild as they seem; they have a finite, predictable structure that can be understood and described completely.

The Story of the Symmetric Shadow

Let's dive into the adventure. Imagine you have a giant, magical grid of numbers. In our everyday world, a grid is just a table with rows and columns. But in this math story, the grid is special: it has a fixed number of rows (let's say $k$ rows), but the number of columns ( $N$ ) can grow as big as you want. As you add more columns, a magical force called the "Symmetric Group" ($Sym(N)$) comes along. This force is like a chaotic but fair DJ who shuffles the columns around. If you swap column 1 and column 2, the whole grid changes, but the rules that describe the grid stay the same. This is what mathematicians call "symmetry."

Now, imagine you have a machine (a "morphism") that takes these $k \times N$ grids and transforms them into something even more complex: multi-dimensional blocks of numbers called "tensors." Think of a tensor as a cube of numbers, or even a hyper-cube, where the size of each side grows with $N$ . The machine is also fair; it respects the DJ's shuffling. If you shuffle the input grid, the output tensor shuffles in a matching way.

The big mystery was: What does the collection of all possible outputs look like? In math, we call this the "image closure." It's like asking, "If I throw a net over all the possible results this machine can produce, what is the shape of the net?" In the infinite world, this shape could be a monster with infinitely many jagged edges, defined by an infinite list of rules. If that were true, we could never fully describe it.

The authors of this paper say: "Hold on! We can prove that this monster is actually a tame cat."

They show that even though the space is infinite, the shape of the output is defined by just a finite number of patterns. Here is the trick: The rules that define the shape don't need to be written out for every single column. Instead, you only need a few "seed" rules. Once you have those, the symmetry of the universe (the DJ) automatically generates all the other rules you need. It's like having a stamp with a single flower on it. You don't need to draw a million flowers; you just stamp the one flower in a million different places. The paper proves that for these specific types of machines, you only need a finite number of "flower stamps" (orbits of equations) to describe the entire infinite shape.

But there is a second, even cooler part to the story. The authors also prove that this shape is "topologically Noetherian." To understand this, imagine you are a treasure hunter looking for hidden rooms inside a castle. You find a room, then you find a smaller room inside it, then an even smaller one inside that. In a chaotic, infinite castle, you might keep finding smaller and smaller rooms forever, never reaching a bottom. But the authors prove that in this specific symmetric castle, this process must stop. No matter how deep you dig, you will eventually hit a floor where there are no more smaller rooms to find. The chain of "smaller rooms" stabilizes. This is a powerful guarantee of order in a world that could easily be chaotic.

How They Did It: The Magic of "Flattening"

How did they prove this? They used a clever technique called "flattening." Imagine you have a 3D cube of numbers. If you look at it from the side, you can "flatten" it into a 2D sheet (a matrix). The authors realized that if you look at these flattened sheets, they have a special property: they have a low "rank." In math-speak, "rank" is a measure of how complex a matrix is. A low-rank matrix is like a simple picture that can be built from just a few basic strokes.

They proved that the outputs of their machine, when flattened, always look like these simple, low-rank pictures. Because they are simple, they are forced to obey a finite set of rules (specifically, rules about the "determinants" of small sub-grids being zero). By showing that the complex, high-dimensional shape is constrained by these simple, flat rules, they could prove that the whole infinite shape is controlled by a finite number of patterns.

They also had to deal with a tricky part: the "diagonal." In a grid, the diagonal is where the row number matches the column number. Sometimes, the rules for the diagonal are different from the rest. The authors showed that even with these diagonal quirks, the "off-diagonal" parts (the rest of the grid) are so restrictive that they force the whole shape to be well-behaved. They even proved a "tensor completion" result: if you have a partial tensor (with some missing diagonal parts) that follows these simple rules, you can always fill in the missing parts to make a full, valid tensor without breaking the rules.

What This Means for the Math World

This paper doesn't just solve a puzzle; it opens a door. It confirms that a specific class of infinite shapes, which appear in fields like algebraic statistics (where they model things like how genes interact or how data is correlated), are actually manageable. Before this, mathematicians knew that some simple shapes were well-behaved, but they weren't sure if the more complex ones (like those involving tensors) would stay under control.

The authors prove that if you start with a "width-1" shape (a simple matrix) and map it to a tensor, the result is always a "finite-pattern" shape. They also prove that the "kernel" (the set of rules that get crushed to zero by the machine) is likely finite, though they admit they haven't fully proved that part yet. They also point out a warning: while the shape is well-behaved in a "reduced" sense (ignoring some weird, non-geometric glitches), it might still have some infinite complexity if you look at the "non-reduced" details (like in characteristic 2, a specific type of math arithmetic). But for the main, visible structure, the chaos is tamed.

In short, this paper tells us that symmetry is a superpower. Even in an infinite universe of variables, if you have enough symmetry, the rules don't have to be infinite. You can describe the whole infinite world with a finite list of instructions, and you can be sure that digging deeper into the structure will always lead to a stopping point. It's a beautiful reminder that order can emerge from the most complex, infinite-looking systems.

Technical Summary: Image Closure of Symmetric Wide-Matrix Varieties

Problem Statement
The paper addresses finiteness properties of rings with countably many variables and infinite-dimensional varieties acted upon by the infinite symmetric group, $\text{Sym}(\infty)$ . While it is established that the polynomial ring in $k \times \mathbb{N}$ variables is $\text{Sym}(\mathbb{N})$ -Noetherian (implying the space of $k \times \mathbb{N}$ -matrices is topologically $\text{Sym}(\mathbb{N})$ -Noetherian), this property fails for higher-dimensional structures. Specifically, the space of $N \times N$ -matrices or $N \times \cdots \times N$ -tensors under the natural $\text{Sym}(\mathbb{N})$ action is not equivariantly Noetherian. This creates a challenge for studying varieties relevant to applications (such as the $k$ -factor model and hierarchical models in algebraic statistics) that naturally reside in these tensor spaces and are preserved by the symmetric group. The central problem is to identify $\text{Sym}(\mathbb{N})$ -stable subvarieties within these tensor spaces that are defined by finitely many orbits of equations and are themselves equivariantly Noetherian.

Methodology
The authors utilize the framework of FI-algebras and FIop-schemes, a language introduced in [CEF15] and developed in [NR19; DEF22], to study structures with increasing numbers of variables acted upon by increasing symmetric groups.

FI-algebras: Covariant functors from the category of finite sets with injective maps ( $\mathbf{FI}$ ) to the category of algebras.
Width: A key concept is the "width" of an FI-algebra. An algebra is "generated in width $\le d$ " if it is generated by elements that depend on at most $d$ variables.
Reduction Strategy: The proof of the main theorem proceeds by reducing the general case to free FI-algebras.
1. The authors first show that for a morphism $\phi: B \to A$ , if the image closure is well-behaved for a surjective pre-image of $B$ , it is well-behaved for $B$ itself. This allows assuming $B$ is a free FI-algebra (specifically, a tensor product of algebras $B_d$ ).
2. Similarly, they reduce the target algebra $A$ to a free algebra generated in width $\le 1$ .
Flattening and Rank: The core technical mechanism involves "flattening" tensors (viewing them as matrices) and analyzing their rank. The authors define subsets $Z_{d,l}$ of the spectrum of $B_d$ where all off-diagonal $(l+1) \times (l+1)$ sub-determinants of flattenings vanish.
Tensor Completion: A crucial intermediate result (Proposition 5.2.10) demonstrates that tensors in $Z_{d,l}$ (off-diagonal tensors with vanishing high-order sub-determinants) can be completed to full tensors of bounded rank. This links the vanishing of equations to the existence of a parameterization by a lower-width algebra.

Key Contributions and Results
The paper establishes the following main theorem regarding morphisms between FI-algebras:

Theorem 1.1.1 (Main Theorem): Let $\phi: B \to A$ be a morphism of FI-algebras over a Noetherian ring $K$ , where $A$ and $B$ are finitely generated, and $A$ is generated in width at most 1. Then:

Finite Definition: The Zariski closure of the image of the dual map $\text{Spec}(A) \to \text{Spec}(B)$ is defined, set-theoretically, by finitely many elements in $B$ . Equivalently, the radical of the kernel of $\phi$ is the radical of a finitely generated FI-ideal.
Topological Noetherianity: The image closure is topologically Noetherian, meaning it satisfies the descending chain condition on reduced closed subschemes.

Corollary 1.1.2: Applied to the geometric setting, for any $\text{Sym}(\mathbb{N})$ -equivariant morphism from the scheme of $k \times \mathbb{N}$ -matrices to the scheme of $N \times \cdots \times N$ -tensors, the image closure is defined by finitely many $\text{Sym}(\mathbb{N})$ -orbits of equations and is topologically $\text{Sym}(\mathbb{N})$ -Noetherian.

Specific Technical Results:

Lemma 5.2.4: The subset $Z_{d,l}$ (defined by vanishing off-diagonal sub-determinants) is defined by finitely many equations in $B_d$ .
Proposition 5.2.10: Tensors in $Z_{d,l}$ can be completed to full tensors of bounded rank $N$ , independent of the size of the underlying set.
Lemma 5.2.11: For any morphism from a free FI-algebra to a width-1 FI-algebra, the image is contained in a closed subset defined by finitely many equations, which is itself contained in the image of a width-1 FIop-scheme.

Significance and Claims
The authors claim that this work provides a robust framework for establishing finiteness properties in infinite-dimensional equivariant geometry.

Unification: The results immediately imply previous significant findings, such as the main result in [Dra10] regarding the Gaussian $k$ -factor model, by showing that the map defining the model fits the width-1 generation criteria.
Set-Theoretic vs. Scheme-Theoretic: The paper explicitly notes that while the image closure is set-theoretically defined by finitely many orbits, it is not necessarily scheme-theoretically Noetherian (i.e., the kernel itself may not be finitely generated, and non-reduced subschemes may form infinite descending chains, particularly in characteristic 2).
Conjecture: The authors state a conjecture (Conjecture 1.1.3) that the FI-ideal $\text{Ker}(\phi)$ is finitely generated, though they do not prove this in the current work.

The paper does not propose new experimental applications but rather solidifies the theoretical foundation for studying varieties in matrix and tensor spaces that arise in algebraic statistics and discrete geometry, confirming that their image closures under equivariant maps possess the desirable property of being defined by finitely many symmetric orbits.

The Story of the Symmetric Shadow

How They Did It: The Magic of "Flattening"

What This Means for the Math World

More like this