🔢 mathematics

The asymptotic in Waring's problem over function fields via singular sets in the circle method

This paper establishes stronger asymptotic results for Waring's problem over function fields than those known for integers by treating minor arcs as complete exponential sums and bounding them via the dimension of singular loci using Katz's results and tangent space calculations.

Original authors: Will Sawin

Published 2026-03-25

📖 5 min read🧠 Deep dive

CC0 1.0

Original authors: Will Sawin

Original paper dedicated to the public domain under CC0 1.0 (http://creativecommons.org/publicdomain/zero/1.0/). ✨ 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: A Cosmic Puzzle Game

Imagine you are playing a game where you have a giant bag of building blocks (polynomials). Your goal is to build a specific, complex structure (a target polynomial) by stacking exactly $s$ smaller towers, where each tower is made by taking a single block and raising it to the power of $k$ (like building a tower of 3 blocks, or 4 blocks, or 5 blocks).

This is Waring's Problem. In the world of regular numbers (integers), mathematicians have been trying to figure out: How many towers ( $s$ ) do I need to guarantee I can build any structure?

For a long time, the answer was "a lot." But recently, mathematicians found a way to do it with fewer towers. This paper takes that same puzzle but moves it from the world of regular numbers to a different universe: Function Fields.

What is a Function Field?
Think of regular numbers as a flat, endless road. Function fields are like a landscape of hills and valleys (curves). Instead of building with numbers, you are building with polynomials (equations like $x^2 + 3x + 1$ ). The rules are similar, but the geometry of the "landscape" changes how the blocks fit together.

The Main Discovery: Doing It with Fewer Blocks

The author, Will Sawin, proves that in this polynomial landscape, you can solve the puzzle with fewer towers ( $s$ ) than was previously thought possible, and even fewer than the best known methods for regular numbers.

The Analogy:
Imagine you are trying to fill a swimming pool with buckets of water.

Old Method: You needed 20 buckets to be sure you could fill any size pool.
New Method (Sawin): By looking at the shape of the pool and the water in a new way, Sawin shows you only need 12 buckets (for certain pool sizes) to guarantee a full pool.

This is a big deal because it means the "rules" of this mathematical universe are more efficient than we thought.

The Secret Weapon: The "Singular Locus"

How did he do it? He didn't use the usual heavy machinery (like complex calculus or "squeezing" the numbers). Instead, he used a geometric trick involving a Singular Locus.

The Metaphor: The Foggy Mountain Pass
Imagine you are trying to walk through a mountain range.

The Smooth Path: Most of the time, the ground is flat and easy to walk on. In math, this is where the "exponential sums" (the calculation of how many ways you can build the structure) behave nicely.
The Singular Locus (The Foggy Pass): This is a specific, tricky spot in the landscape where the ground gets bumpy, steep, or confusing. It's where the math gets "singular" (weird).

In the past, mathematicians tried to calculate the difficulty of the whole mountain by looking at the smooth parts and guessing about the foggy parts. Sawin realized: "Don't guess. Measure the fog."

He treated the "foggy pass" (the singular locus) not as a problem to be avoided, but as a specific shape to be measured. He used a powerful tool from a mathematician named Katz (think of Katz as a master cartographer) to measure exactly how "big" and "complex" this foggy area is.

The Breakthrough:
Sawin found that if you can measure the size of this "foggy pass" accurately, you can prove that the rest of the mountain is actually very easy to cross. The smaller the foggy area, the fewer buckets (towers) you need to solve the puzzle.

Why This Matters: The "Manin's Conjecture" Connection

The paper also connects this to Manin's Conjecture, which is like a prediction about how many "paths" exist between two points in this mathematical landscape.

The Fermat Hypersurface: This is a specific, famous shape in the landscape (like a perfect sphere or a specific type of crystal).
The Result: Sawin's method allows mathematicians to count the number of paths on these shapes much more accurately. It's like having a GPS that finally tells you the exact number of roads leading to a specific city, rather than just an estimate.

The "Pugin" Connection and the "Gap"

The paper mentions a previous attempt by a mathematician named Pugin who tried a similar trick.

The Analogy: Pugin tried to build a bridge over the foggy pass, but he made a small error in his blueprint (a "gap" in the logic). He assumed the bridge would be shorter than it actually was.
Sawin's Fix: Sawin went back, fixed the blueprint, and showed exactly how to calculate the bridge's length using "tangent spaces" (which are like measuring the slope of the ground at a single point). By breaking the problem down into tiny, manageable slices (stratification), he proved the bridge is stable and the math works.

Summary for the Non-Mathematician

The Problem: How many "power towers" do you need to build any polynomial structure?
The Old Answer: You need a lot.
The New Answer: You need fewer, thanks to a new way of looking at the geometry of the problem.
The Method: Instead of fighting the "weird" parts of the math, the author measured them precisely. He realized that the "weirdness" is actually very small and contained, which makes the whole system much more efficient.
The Result: We now have better rules for counting solutions in the world of polynomials, which helps us understand the deep structure of numbers and shapes in a way that was previously impossible.

In short: Sawin found a shortcut through the mathematical wilderness by mapping the "danger zones" so precisely that he realized the path is actually much shorter than anyone thought.

1. Problem Statement

The paper addresses Waring's problem over function fields, specifically seeking the asymptotic formula for the number of solutions to the equation:
$\sum_{i=1}^s a_i^k = f$
where $a_i$ are polynomials in $\mathbb{F}_q[T]$ of degree at most $e$ , and $f$ is a polynomial of degree at most $ke$.

Key Objectives:

To establish an asymptotic formula of the form:
$\#\{a \in \mathbb{F}_q[T]^s \mid \deg a_i \le e, \sum a_i^k = f\} = (1+o(1)) \mathcal{M}$
where $\mathcal{M}$ is the expected main term involving singular series and local factors.
To determine the minimum number of variables $s$ (as a function of the degree $k$ ) required for this asymptotic to hold.
To extend these results to Manin's conjecture for Fermat hypersurfaces over function fields, counting rational maps from a curve $C$ to the hypersurface.

Context:
Over the integers, the best known bounds for $s$ rely on the Vinogradov Mean Value Theorem (VMVT) and the breakthrough work of Bourgain, Demeter, and Guth. However, these bounds are roughly quadratic in $k$ (e.g., $s \approx k^2$ ). Sawin aims to achieve linear bounds in $k$ (e.g., $s \approx 4k$ ) in the function field setting, which would represent a significant improvement over the integer case if the parameters allow.

2. Methodology

The paper employs the Circle Method, but with a distinct geometric approach to handling the "minor arcs," diverging from traditional analytic number theory techniques used over integers.

A. The Circle Method Setup

The problem is transformed into an exponential sum over the dual space of sections of a line bundle $L$ on a curve $C$ . The sum is split into:

Major Arcs: Corresponding to rational approximations with small denominators (low-degree subschemes). These are handled using standard analytic techniques, evaluating local factors and singular series.
Minor Arcs: Corresponding to "irrational" approximations.

B. The Novel Approach to Minor Arcs

Instead of using Weyl differencing or efficient congruencing (standard analytic tools), Sawin treats the minor arc sums as complete exponential sums over finite fields.

Geometric Interpretation: The set of polynomials of degree $\le e$ is viewed as an affine space (or vector space) over $\mathbb{F}_q$ . The exponential sum is associated with a polynomial function on this space.
Katz's Estimate: The author applies a theorem by Nicholas Katz which bounds exponential sums in terms of the dimension of the singular locus of the underlying polynomial.
- The bound is of the form $|S(\alpha)| \ll q^{(N + \dim(\text{Sing})) / 2}$ , where $N$ is the dimension of the space.
- To get a strong bound (power saving), one must prove that the dimension of the singular locus is small.

C. Analyzing the Singular Locus

The core technical innovation lies in the precise estimation of the dimension of the singular locus for the Fermat polynomial $F = \sum x_i^k$ .

Characterization: The singular locus is defined by the vanishing of the derivative of the exponential sum phase. For the Fermat case, this leads to a system of equations involving the $(k-1)$ -th powers of the variables.
Stratification: The author stratifies the singular locus based on the joint root multiplicities of the polynomials involved.
Tangent Space Calculations: Instead of attempting to calculate the dimension directly (which Pugin previously attempted with a gap), Sawin estimates the dimension by bounding the dimension of the tangent space at arbitrary points.
- A key simplification involves taking the logarithmic derivative, which transforms the $(k-1)$ -th power equation into a linear multiplication by $k-1$ .
- This allows for a rigorous bound on the dimension of the solution space, even in cases with repeated roots.
Convex Hull Argument: The dimension bound is expressed in terms of a convex set $\Delta_{k,p}$ defined by the condition $p \mid (k-1)i - j$ . The parameter $\gamma_{k,p}$ (related to the dilation of this set) determines the final bound on $s$ .

3. Key Contributions and Results

A. Improved Asymptotic Bounds

The paper proves that for a finite field $\mathbb{F}_q$ of characteristic $p > k$ , the asymptotic holds provided $s$ satisfies:
$s > \max\left( \frac{2(k - \gamma_{k,p} - 2L_{k,q})}{1 - \gamma_{k,p} - 2L_{k,q}}, \frac{2k}{1 - 2(k-1)L_{k,q}} \right)$
where $L_{k,q} = \frac{\log(k+1)}{\log q}$ .

Significance of the Bound:

For sufficiently large $q$ and $p$ , the bound simplifies to $s > 4k - 6 + \frac{4}{k}$ .
This implies $s \ge 4k - 5$ for $k \ge 4$ .
Comparison: This is a linear dependence on $k$ $k$ , significantly improving upon the quadratic dependence ( $s \approx k^2$ $s \approx k^{2}$ ) known for integers via VMVT.
- Example: For $k=4$ , integers require $s \ge 15$ ; this paper requires $s \ge 12$ .
- Example: For $k=5$ , integers require $s \ge 23$ ; this paper requires $s \ge 15$ .

B. Generalization to Manin's Conjecture

The results are applied to Manin's conjecture for Fermat hypersurfaces $X \subset \mathbb{P}^n$ defined by $\sum x_i^d = 0$ .

The paper derives an asymptotic formula for the number of rational maps $f: C \to X$ of degree $e$ .
The formula matches the expected form involving the Tamagawa number $\tau_C(X)$ and the height function, confirming the conjecture for Fermat hypersurfaces over function fields under the same linear conditions on the number of variables.

C. Correction of Previous Work

The author identifies and corrects a gap in a previous paper by Pugin regarding the dimension calculation of the singular locus. Pugin's argument relied on an incorrect assumption about the dimension of inverse images under diagonal morphisms. Sawin's explicit stratification and tangent space analysis provide a rigorous foundation.

D. Speculative Extension to General Hypersurfaces

In the final section, the author outlines a potential approach for arbitrary smooth hypersurfaces (not just Fermat).

The singular locus is related to the moduli space of maps from the curve $C$ to a specific blowup of $\mathbb{P}^n \times \mathbb{P}^n$ .
While the dimension of these moduli spaces is not currently known precisely, the paper suggests that if upper bounds for these moduli spaces can be established, the method could extend to general hypersurfaces.

4. Significance

Breaking the Quadratic Barrier: The paper demonstrates that the function field setting allows for much stronger results (linear in $k$ ) than the integer setting, highlighting the power of geometric methods (algebraic geometry over finite fields) in number theory.
Geometric Circle Method: It solidifies the "geometric" approach to the Circle Method, where minor arcs are treated as complete exponential sums and bounded via cohomological dimension estimates (Katz's work) rather than analytic oscillation estimates.
Manin's Conjecture: It provides the first unconditional asymptotic results for Manin's conjecture on Fermat hypersurfaces over function fields with a linear dependence on the degree, a regime previously inaccessible.
Technical Rigor: The explicit calculation of the singular locus dimension via stratification and logarithmic derivatives sets a new standard for handling exponential sums in high-dimensional vector spaces over finite fields.

In summary, Will Sawin's paper leverages deep results from algebraic geometry (specifically the cohomology of exponential sums and the geometry of singular loci) to solve a classical additive number theory problem with bounds that surpass current knowledge in the integer case.