🔢 mathematics

Parity of $k$ -differentials in genus zero and one

This paper establishes the spin parity of $k$ -differentials on Riemann surfaces of genus zero and one by proving a previously conditional number-theoretic hypothesis through a reformulation involving Jacobi symbols and a combinatorial identity verified by the AxiomProver system and formalized in Lean.

Original authors: Dawei Chen, Evan Chen, Kenny Lau, Ken Ono, Jujian Zhang

Published 2026-07-07

📖 4 min read🧠 Deep dive

CC BY 4.0

Original authors: Dawei Chen, Evan Chen, Kenny Lau, Ken Ono, Jujian Zhang

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: Solving a Mathematical Puzzle

Imagine you have a collection of flexible, rubbery sheets (mathematicians call these "Riemann surfaces"). On these sheets, you can draw special patterns called k-differentials. Think of these patterns like a grid of roads that guide a car. Sometimes the roads are smooth, and sometimes they have sharp turns or dead ends (called "zeros" and "poles").

Mathematicians want to organize all these possible patterns into groups. They know that some patterns can be smoothly transformed into one another, while others are stuck in a separate "island" and can never be reached from the first group. The goal of this paper is to figure out exactly how many islands there are and which patterns belong to which island.

The Specific Problem: The "Spin" of the Pattern

For a long time, mathematicians knew how to sort these patterns for simple cases (like when the roads are perfectly smooth or when the sheet is a simple sphere). However, for more complex cases involving odd numbers of turns (called "odd k"), they hit a wall.

They discovered a special property called Spin Parity.

The Analogy: Imagine every pattern has a hidden "handedness," like a left hand or a right hand. If you try to turn a left-handed pattern into a right-handed one without tearing the sheet, you can't.
The Mystery: For simple sheets (genus 0 and 1), the authors knew how to calculate this handedness, but their calculation relied on a guess. They had a hunch that a specific number-counting rule was always true, but they couldn't prove it. Without proving this rule, their whole classification system was "conditional" (it only worked if the guess was right).

The Guess: A Number Game

The unproven guess (Conjecture 1.1) was a game involving counting pairs of numbers.

The Game: Pick an odd number $k$ . Count how many pairs of smaller numbers $(b_1, b_2)$ fit a specific set of rules involving addition and division.
The Prediction: The authors guessed that the total count of these pairs would always be either "even" or "odd" in a very specific way, depending only on the number $k$ .
The Stakes: If this guess is true, the "handedness" of all those complex patterns on the sheets is solved. If it's false, the classification falls apart.

The Breakthrough: The AI Detective

This is where the paper gets unique. The authors didn't just sit down and solve this with a pencil and paper. They used a new, experimental AI system called AxiomProver.

The Reformulation: The human authors realized that this messy counting game could be rewritten using a standard mathematical tool called a Jacobi Symbol (think of this as a special "parity switch" used in number theory).
The AI's Job: They fed this reformulated problem to AxiomProver. The AI didn't just guess the answer; it acted like a rigorous logic detective. It found a hidden connection between the counting game and a known combinatorial identity (a rule about how numbers add up).
The Proof: The AI proved that the counting game always results in the parity the authors predicted. It then wrote a formal, computer-verifiable proof of this fact in a language called Lean.

The Result: The Puzzle is Solved

Because the AI proved the guess was correct, the "conditional" results in the paper become absolute facts.

What they found: They now have a complete, unshakeable rule for determining the "handedness" (spin parity) of these patterns on simple sheets (genus 0 and 1) for any odd number of turns.
The Formula: They provided a simple formula to calculate this. You look at the "sizes" of the turns (zeros and poles), check them against the prime factors of your number $k$ , and count how many don't match a specific pattern. That count tells you the handedness.

Summary

In short, this paper is about:

The Problem: Classifying complex geometric patterns on simple shapes.
The Blockage: A missing proof for a number-theory guess that was holding up the whole theory.
The Solution: An AI system (AxiomProver) that translated the guess into a standard math language, found the logical proof, and verified it with computer code.
The Outcome: The classification of these patterns is now complete and proven true, removing all doubt.

Note: The paper explicitly states that the formal proof was done on the combinatorial identity (the number game), not the geometric shapes themselves. The geometry part was already understood; the number game was the missing key.

Technical Summary: Parity of k-Differentials in Genus Zero and One

Problem Statement
The paper addresses the classification of connected components of the moduli space $\Omega^k\mathcal{M}_g(\mu)$ , which parameterizes primitive $k$ -differentials on Riemann surfaces of genus $g$ with prescribed zero and pole orders $\mu = (m_1, \dots, m_n)$ . While the classification of connected components is well-established for $k=1$ (differentials) and $k=2$ (quadratic differentials), the case for $k \ge 3$ remains largely open.

A primary invariant used to distinguish these components is spin parity. For $k$ -differentials where the orders are even (parity type), the spin parity is defined via a canonical cyclic cover $\pi: \hat{X} \to X$ of degree $k$ , where the pullback of the $k$ -differential becomes the $k$ -th power of a 1-form $\hat{\omega}$ . The spin parity of the original $k$ -differential is the spin parity of the pair $(\hat{X}, \hat{\omega})$ .

For even $k$ , the spin parity is uniform across the moduli space and determined explicitly by $\mu$ . However, for odd $k$ , the spin parity can vary within the moduli space, making its determination nontrivial. Previous work by Chen and Gendron [6] provided explicit formulas for the spin parity in genus zero and one, but these results were conditional on a specific number-theoretic conjecture (Conjecture A.10 in [6]).

Methodology
The authors resolve the conditional nature of the previous results by proving the underlying number-theoretic conjecture. The methodology proceeds in two main stages:

Reformulation via Jacobi Symbols:
The core of the proof involves translating the combinatorial definition of the conjecture into the language of Jacobi symbols. The conjecture concerns $N_k(n)$ , the number of pairs of positive integers $(b_1, b_2)$ satisfying specific bounds and a congruence condition $b_2 \equiv n b_1 \pmod k$ .
The authors observe that the parity of the term $\lfloor \frac{k+1}{4} \rfloor$ appearing in the conjecture is periodic modulo 8 and corresponds to the value of the Jacobi symbol $\left(\frac{2}{k}\right)$ .
Combinatorial Reduction:
The authors define a combinatorial sum $F_k(a) = \sum_{i=1}^{(k-1)/2} \lfloor \frac{ai + (k-1)/2}{k} \rfloor$ . Using standard properties of Jacobi symbols (specifically Eisenstein's Lemma and Gauss's Lemma), they establish a relationship between $F_k(a)$ and the parity of $a$ :
- If $a$ is odd, $F_k(a) \equiv 0 \pmod 2$ .
- If $a$ is even, $F_k(a) \equiv \lfloor \frac{k+1}{4} \rfloor \pmod 2$ .
They then prove a key combinatorial identity (Lemma 2.5):
$N_k(n) = F_k(n+1) - F_k(n)$
This identity reduces the problem to analyzing the parity of $F_k(n)$ and $F_k(n+1)$ . Since $n$ and $n+1$ have opposite parity, and the hypothesis requires $\gcd(n, k) = \gcd(n+1, k) = 1$ , the result follows directly from the properties of $F_k$ .

Key Contributions and Results

Proof of Conjecture 1.1: The paper provides a complete proof of Conjecture 1.1 (originally from [6]), which states that for an odd integer $k \ge 3$ and an integer $n$ coprime to both $k$ and $k+1$ :
$N_k(n) \equiv \left\lfloor \frac{k+1}{4} \right\rfloor \pmod 2$
Unconditional Classification: By proving the conjecture, the authors render the spin parity formulas for genus zero and one unconditional.
- Genus 0: The spin parity of $\Omega^k\mathcal{M}_0(2\mu)$ is determined by $n_k(\mu) \pmod 2$ , where $n_k(\mu)$ counts indices $i$ where the $Q$ -adic valuation of $m_i$ differs in parity from that of $k$ .
- Genus 1: The spin parity of the connected component with rotation number $d$ is determined by $n_k(\mu) + d + 1 \pmod 2$ .
Alternative Formulation: The paper offers a more concise definition for the invariant $n_k(\mu)$ using Jacobi symbols:
$n_k(\mu) = \# \left\{ i : \left(\frac{2}{\gcd(k, m_i)}\right) \neq \left(\frac{2}{k}\right) \right\}$

Significance and Claims
The paper claims that its primary contribution is the proof of the number-theoretic hypothesis that was previously a barrier to the complete classification of spin parity for $k$ -differentials in low genus. The geometric results regarding $k$ -differentials themselves were established in prior work [6]; this paper removes the conditional assumption, thereby "completely determining the spin parity of $k$ -differentials in genus zero and one for all $k$ ."

The authors explicitly note that the proof of the combinatorial identity (Lemma 2.5) and the main theorem were autonomously discovered by AxiomProver, an AI system under development. The formal verification of the combinatorial identity was performed in Lean/Mathlib. The paper emphasizes that the formalization applies to the combinatorial identity, not the geometric results on $k$ -differentials.

The work serves as a test case for AxiomProver, demonstrating its capability to discover non-trivial connections between combinatorial sums and Jacobi symbols, and to generate formal proofs for complex number-theoretic identities.