Functionality for isomorphism classes of curves and hypersurfaces

This paper presents algorithms grounded in invariant theory to address geometric problems concerning curves and hypersurfaces, with a primary focus on those of genus 2, 3, and 4, while also incorporating new theoretical results derived from the first author's PhD thesis.

The Gist

Imagine you are a master architect who designs unique, intricate buildings (mathematical curves). You have a massive library of blueprints, but you've lost the labels. All you have left are a few specific measurements of the building's features—like the height of the tallest tower, the width of the foundation, or the angle of the roof. These measurements are called invariants.

This paper is essentially a new, super-charged toolkit for architects (mathematicians) that helps them do three main things:

1. Identify a building just by looking at its measurements.
2. Rebuild the entire blueprint from just those measurements.
3. Prove if two different-looking blueprints are actually the same building, just rotated or viewed from a different angle.

Here is a breakdown of the paper's "magic tricks" using everyday analogies:

1. The "Fingerprint" Problem (Invariants)

In the world of math, curves (like loops or twisted shapes) can look very different depending on how you draw them. One might look like a squashed circle, another like a wavy line, but they might actually be the same shape, just stretched or rotated.

• The Old Way: To tell if two shapes are the same, you'd have to try to physically rotate and stretch one until it matched the other. This is slow and hard.
• The New Way (Invariants): Think of invariants as a fingerprint. No matter how you rotate or stretch the shape, the fingerprint stays the same.
  • The authors have updated the "fingerprint database" for complex shapes (specifically curves with 2, 3, or 4 "holes" or loops).
  • They figured out how to calculate these fingerprints even when the math gets weird (like in "positive characteristic," which is a fancy way of saying the math rules change slightly, similar to how a game might have different rules on a different planet).

2. The "Reverse Engineering" Problem (Reconstruction)

Imagine you have a photo of a building's shadow (the invariants) and you need to build the actual house (the curve) from scratch.

• The Challenge: Many different houses can cast the same shadow. How do you know which one to build?
• The Solution: The authors developed a new 3D printer algorithm.
  • For simple shapes (Genus 2 and 3), they have a reliable printer that works almost every time.
  • For the most complex shapes (Genus 4), they created a brand-new, highly sophisticated printing method. It's like having a recipe that says, "If you mix these specific measurements together in this exact order, you get the original building."
  • The "Generic" Trick: They realized that for most random shapes, there's a shortcut. You don't need to solve a massive puzzle; you just need to find a specific "key" (a mathematical tool called a contravariant) that unlocks the shape directly.

3. The "Body Double" Problem (Isomorphisms)

Sometimes you have two blueprints, and you need to know: "Are these the same building?"

• The Tricky Part: If a building has perfect symmetry (like a square), it can be rotated in many ways and still look the same. This makes it hard to tell if two blueprints are identical or just different versions of the same thing.
• The New Method:
  • For Symmetry-Free Buildings: The authors found a lightning-fast way to check. It's like checking if two unique snowflakes are the same by comparing their specific crystal patterns. If the patterns match perfectly, the buildings are identical.
  • For Symmetrical Buildings: If the building has symmetry (like a perfect sphere), the math gets harder. The authors created a "detective" method that uses a mix of fast checks and heavy-duty math (called Gröbner bases) to solve the puzzle. They even figured out how to handle the "Genus 4" cases, which were previously too difficult to solve efficiently.

4. The "Twist" Problem

In math, you can sometimes take a shape and "twist" it (like twisting a rubber band) so it looks different but is fundamentally the same object.

• The paper provides tools to list all the possible "twisted" versions of a shape. This is useful for cryptography and understanding how shapes behave in different mathematical universes.

Why Does This Matter?

Think of this paper as updating the operating system for a super-computer that designs the universe's geometry.

• Before: Mathematicians had to write custom code for every new shape, and it often crashed or took years to run.
• Now: They have a unified, robust software package (available in a tool called Magma) that can instantly identify, rebuild, and compare these complex shapes.

The Bottom Line:
The authors (Bouchet, Lerici, Sijssling, and Ritzenthaler) have taken a very abstract, difficult branch of math and turned it into a set of practical, automated tools. They solved the "Genus 4" problem (the hardest level of the game) and made the tools work faster and more reliably, even in tricky mathematical environments. It's like upgrading from a hand-drawn map to a GPS that works in any weather.

Technical Summary

Here is a detailed technical summary of the paper "Functionality for Isomorphism Classes of Curves and Hypersurfaces" by Bouchet, Lercier, Sijsling, and Ritzenthaler.

1. Problem Statement

The paper addresses the computational challenges in algebraic geometry regarding the classification, reconstruction, and isomorphism testing of algebraic curves, specifically focusing on hyperelliptic curves and non-hyperelliptic curves of genus 2, 3, and 4.

The core problems tackled are:

• Classification: Determining when two curves are isomorphic using invariant theory.
• Reconstruction: Given a set of invariants (which characterize an isomorphism class), constructing an explicit equation (a model) for a curve in that class, ideally defined over the field generated by those invariants.
• Isomorphism Computation: Explicitly computing the linear transformations (isomorphisms) between two given curves.
• Twists and Automorphisms: Computing the automorphism groups of curves and determining their twists (forms of the curve defined over a base field that become isomorphic over an algebraic closure).

While Magma (a computer algebra system) had existing functionality for these tasks, it was fragmented, limited to specific characteristics, or inefficient for generic cases. This paper unifies these efforts, introduces new theoretical results, and provides optimized algorithms.

2. Methodology

The authors employ a combination of classical invariant theory, modern computational algebra, and algorithmic geometry.

A. Invariant Theory

• Hyperelliptic Curves: Utilize the theory of binary forms. The isomorphism class of a genus $g$ hyperelliptic curve corresponds to the orbit of a binary form of degree $2g+2$(or$2g+1$) under the action of$GL_2$. The paper discusses generators for the invariant rings$R_g(k)$, addressing the complexities of positive characteristic (e.g.,$p=2, 3, 5, 7$).
• Non-Hyperelliptic Curves:
  • Genus 3: Represented as smooth plane quartics (ternary forms of degree 4). The authors utilize the Dixmier–Ohno invariants (13 generators).
  • Genus 4: Represented as the intersection of a quadric and a cubic surface. The authors use invariants derived from the first author's PhD thesis, distinguishing between curves lying on smooth quadrics and those on quadric cones.

B. Reconstruction Algorithms

The paper introduces a general framework for reconstruction from invariants, extending Mestre's method for binary forms.

• Contravariants: The core technique involves constructing linear bases of contravariants (polynomials in the coefficients of the form and the variables) to build a system of equations.
• Taylor-like Formula: By evaluating these contravariants, the authors construct a collection of polynomials in a larger ambient space (via Veronese embedding).
• Generic Recovery: For generic curves (where the automorphism group is trivial), this process yields a single polynomial explicitly defined by the invariants, providing a birational inverse to the quotient map.
• Specific Cases:
  • Hyperelliptic: Uses Mestre's method involving the intersection of a conic and a plane curve.
  • Plane Quartics: Uses specific contravariants of order 1 and 2.
  • Genus 4: Uses a novel approach handling the redundancy of the cubic surface equation (adding a multiple of the quadric).

C. Isomorphism Computation

• Generic Hypersurfaces: A new strategy uses covariants to reduce the isomorphism problem to solving a system of monomial equations. If covariants evaluated at two forms $f$ and $g$ form a basis, the isomorphism matrix can be recovered directly.
• Automorphism Detection: The method provides a criterion (Proposition 2.1.3) to detect if a hypersurface has a trivial automorphism group.
• Non-Generic Cases: When covariants fail (e.g., non-trivial automorphisms), the authors resort to Gröbner basis computations or specialized models based on the automorphism strata.
• Hyperelliptic Specifics: Leverages the correspondence with binary forms to compute isomorphisms via $GL_2$ actions.

3. Key Contributions

Theoretical Advances

1. Genus 4 Invariants: The paper details the first systematic construction of invariants for genus 4 curves (both smooth and singular quadric cases), based on the first author's PhD thesis.
2. Positive Characteristic Validity:
   • Proves that the Dixmier–Ohno invariants remain generators for the invariant ring of ternary quartics in characteristic $p > 13$ (Appendix A).
   • Provides separating invariants for hyperelliptic curves in characteristics 2, 3, 5, and 7, where classical generators often fail.
3. Reconstruction for Genus 4: Presents a complete explicit reconstruction algorithm for generic genus 4 curves, a significant extension of previous work limited to genus 2 and 3.

Algorithmic and Implementation Contributions

1. Unified Magma Packages: The authors have gathered scattered functions into coherent Magma packages:
   • HyperellipticCurveFromIgusaInvariants (Genus 2).
   • HyperellipticCurveFromShiodaInvariants (Genus 3).
   • HyperellipticCurveFromBrouwerPopoviciuInvariants (Genus 4).
   • PlaneQuarticFromDixmierOhnoInvariants (Genus 3).
   • Genus4CurveFromInvariants (Genus 4).
2. Efficient Isomorphism Detection:
   • A new covariant-based method for generic hypersurfaces that is significantly faster than traditional Gröbner basis approaches.
   • A criterion to detect trivial automorphism groups, allowing for immediate isomorphism computation without solving complex systems.
3. Optimization:
   • Implementation of "variation of conics" to speed up finding rational points for hyperelliptic reconstruction.
   • Reduction of coefficient sizes in reconstruction outputs over $\mathbb{Q}$ using Stoll's methods.

4. Results

• Reconstruction: The algorithms successfully reconstruct generic curves of genus 2, 3, and 4 from their invariants. For genus 4, this is a novel result.
• Isomorphism Speed: The new covariant-based isomorphism routine is orders of magnitude faster than native Magma routines for plane quartics with trivial automorphism groups (e.g., 0.14s vs 3.20s in the provided example).
• Characteristic Handling: The code handles a wide range of characteristics, including specific adaptations for $p=2, 3, 5, 7$ where standard invariant sets are insufficient.
• Twists: The paper provides functional tools to compute twists of curves over finite fields, utilizing the computed automorphism groups.

5. Significance

• Practical Utility: This work serves as a "vademecum" (handbook) for researchers using Magma to study curves of genus 2–4. It consolidates previously scattered research into a usable software library.
• Theoretical Depth: It bridges the gap between classical 19th-century invariant theory and modern computational algebra, resolving open questions about the generation of invariant rings in positive characteristic.
• Completeness: It provides the first complete explicit reconstruction algorithm for genus 4 curves, a class of objects previously difficult to handle computationally.
• Future Directions: The paper explicitly outlines open problems, such as reconstructing curves with non-trivial automorphism groups in genus 4 and finding generators for invariant rings in small positive characteristics (e.g., $p=2, 3$), setting the agenda for future research in computational algebraic geometry.

In summary, the paper represents a major step forward in the computational arithmetic of algebraic curves, providing robust, efficient, and theoretically grounded tools for classifying and reconstructing curves of genus up to 4 across various field characteristics.