🔢 mathematics

Yoga for Fourier--Mukai partnership

This paper establishes a framework of local algebra and fiber analysis to test for derived equivalences and fully faithful integral transforms under base change, thereby generalizing Orlov's results to singular varieties over arbitrary fields and providing new insights into arithmetic fibrations.

Original authors: Elías Guisado Villalgordo, Pat Lank, Kabeer Manali Rahul, Nebojsa Pavic

Published 2026-03-27

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Elías Guisado Villalgordo, Pat Lank, Kabeer Manali Rahul, Nebojsa Pavic

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 an architect trying to understand two different buildings. You want to know if they are essentially the same structure, just built with different materials or on different plots of land. In the world of advanced mathematics (specifically algebraic geometry), these "buildings" are called schemes, and the "materials" are complex algebraic structures.

This paper, titled "Yoga for Fourier–Mukai Partnership," is like a master guidebook for architects who want to compare these buildings, even when the buildings are a bit broken, cracked, or built on tricky soil (what mathematicians call "singularities").

Here is the breakdown of their work using simple analogies:

1. The Core Concept: The "Magic Mirror" (Integral Transforms)

Imagine you have a magical mirror (called an Integral Transform). If you stand in front of it, it doesn't just show your reflection; it shows you a completely different version of yourself, perhaps as a different person or in a different era.

In math, this mirror takes the "blueprint" of one building (a derived category) and transforms it into the blueprint of another. If the mirror works perfectly, the two buildings are called Fourier–Mukai Partners. They might look totally different on the outside, but their internal logic is identical.

2. The Problem: Broken Buildings and Bad Soil

For a long time, mathematicians could only use this magic mirror on perfect, smooth buildings. If a building had a crack in the wall (a singularity) or was built on unstable ground, the mirror would break, or the reflection would be distorted.

Furthermore, most previous rules only worked if the buildings were in a specific, perfect neighborhood (like an algebraically closed field). If you moved the building to a different country (a different field, like the rational numbers or a finite field), the rules stopped working.

3. The Solution: The "Base Change" Test

The authors of this paper developed a new set of rules called "Yoga" (a term they use to mean a flexible, spiritual practice of moving between different states).

Their big idea is Base Change. Think of this as taking a photo of the building and then zooming in on a specific detail, or looking at the building through a different colored lens.

The Old Way: You had to check the entire building to see if the mirror worked.
The New Way: The authors say, "You don't need to check the whole building! Just look at the special fibers."

The "Special Fiber" Analogy:
Imagine the building is a loaf of bread rising in an oven.

The Base is the oven temperature.
The Fiber is a single slice of bread cut at a specific moment.
The authors discovered that if you check the "slices" (the fibers) at the very bottom of the loaf (the special points), you can predict how the whole loaf behaves. If the mirror works on a tiny slice of bread, it works for the whole loaf, even if the loaf is burnt or crumbly (singular).

4. The "Yoga" Rules (The Main Findings)

The paper establishes a set of "Yoga" rules that allow mathematicians to test if two buildings are partners without having to rebuild them first.

Rule 1: The "Local Algebra" Test. You can test the mirror's power by looking at the building through a microscope (local algebra). If the mirror works on the microscopic level, it works on the macroscopic level.
Rule 2: The "Field Extension" Test. If two buildings are partners in one country, they remain partners even if you move them to a different country (a different field), provided you check the right "slices" first. This is a huge leap because it allows mathematicians to study these structures in "arithmetic" settings (like number theory), not just in smooth, perfect geometric worlds.
Rule 3: The "Smoothness" Invariance. If Building A is perfectly smooth and Building B is its partner, then Building B must also be smooth. You can't have a smooth building paired with a broken one. This is true even if the buildings are in weird, imperfect environments.

5. Why This Matters (The "So What?")

Why should a general audience care about magic mirrors and broken bread?

Solving Puzzles in Number Theory: This helps mathematicians understand the deep connections between shapes and numbers, even when those numbers behave strangely (like in prime number systems).
Understanding the Universe: In physics, the shape of the universe is often described by these complex geometric structures. If we can prove two different shapes are actually "partners," it means they might describe the same physical reality.
Robustness: The authors' methods are "robust." They work even when things are messy, broken, or singular. This is crucial because, in the real world (and in complex math), things are rarely perfect.

Summary

Think of this paper as a universal translator for broken, complex structures. Before this, if a structure was cracked or built on weird soil, we couldn't tell if it was the same as another structure. Now, the authors have given us a "Yoga" practice: look at the tiny, specific details (the fibers), and if those match up, the whole structures are partners, no matter how messy they look.

They have taken a tool that only worked in a pristine, perfect laboratory and made it work in the messy, real world of mathematics.

1. Problem Statement and Context

The paper addresses a fundamental challenge in algebraic geometry: extending the theory of Fourier–Mukai (FM) transforms and derived equivalences from smooth varieties to singular varieties and arithmetic settings (schemes over arbitrary fields, including imperfect fields and discrete valuation rings).

Background: Integral transforms (FM transforms) are exact functors between derived categories of coherent sheaves, defined by a kernel object $K$ . Two schemes $Y_1$ and $Y_2$ are FM partners if such a transform is an equivalence.
The Gap: Classical results (e.g., by Orlov, Bondal, Huybrechts, Stellari) largely rely on smoothness assumptions. In singular settings, bounded coherent complexes ( $D^b_{coh}$ ) are not necessarily perfect ($Perf$), complicating the behavior of adjoints and the preservation of properties under base change.
Specific Question: How do properties like full faithfulness and equivalence of an integral transform behave under base change? Specifically, can one test if a transform is an equivalence by checking its behavior on fibers (special fibers) or after field extensions, even when the total space is singular?

2. Methodology

The authors develop a "yoga" (a systematic framework) combining local algebra, derived category theory, and base change techniques.

Approximable Triangulated Categories: They utilize the framework of Neeman regarding approximable triangulated categories. This allows them to reduce questions about equivalences on $D_{qc}$ (quasi-coherent sheaves) to the behavior on smaller subcategories like $Perf$ (perfect complexes) and $D^b_{coh}$ (bounded coherent complexes).
Relative Perfectness: A central technical tool is the notion of an object being relatively perfect over a base. An object $K$ is relatively perfect over $Y_1$ if the pushforward of $K \otimes Perf(Y_1 \times_S Y_2)$ lands in $Perf(Y_1)$ . This condition ensures the integral transform restricts to the appropriate subcategories.
Base Change and Adjunctions: The authors rigorously analyze how the adjoints (left and right) of an integral transform behave under base change. They construct explicit natural isomorphisms (comparison transformations) between the base-changed transform and the transform of the base-changed kernel.
Descent Arguments: They employ faithfully flat descent and affine morphism techniques to lift properties from fibers (or field extensions) back to the total scheme.

3. Key Contributions and Results

A. Preservation of Adjointness (Theorem 1.1 / Theorem 3.6)

The paper establishes that the property of a kernel $K$ being relatively perfect is local and can be tested on fibers.

Result: Let $S$ $S$ be a Noetherian scheme and $Y_1, Y_2$ $Y_{1}, Y_{2}$ proper flat $S$ $S$ -schemes. For a kernel $K \in D^b_{coh}(Y_1 \times_S Y_2)$ $K \in D_{co h}^{b} (Y_{1} \times_{S} Y_{2})$ , the following are equivalent:
1. $K$ is relatively perfect over $Y_1$ (or $Y_2$ ).
2. The derived pullback $L(t')^*K$ is relatively perfect over the fiber product for any affine morphism $t: T \to S$ .
3. It suffices to check this condition for morphisms from fields ( $Spec(k) \to S$ ) or closed points ( $Spec(\kappa(p)) \to S$ ).
Significance: This generalizes previous results by allowing arbitrary base fields and singular varieties. It confirms that the existence of adjoints is preserved under base change if the kernel is relatively perfect.

B. Descent of Full Faithfulness and Equivalence (Theorem 1.2 / Theorem 4.13)

This is the core theoretical breakthrough. The authors prove that full faithfulness and equivalence are local properties with respect to the base scheme.

Result: Let $K$ $K$ be relatively perfect over $Y_1$ $Y_{1}$ and $Y_2$ $Y_{2}$ . The integral transform $\Phi_K$ $Φ_{K}$ is fully faithful (or an equivalence) on $D^b_{coh}$ $D_{co h}^{b}$ if and only if the induced transform on the fibers $\Phi_{L(t')^*K}$ $Φ_{L (t^{'})^{*} K}$ is fully faithful (or an equivalence) for:
- Any affine morphism $t$ .
- Any morphism from a field $Spec(k) \to S$ .
- Any closed point $Spec(\kappa(p)) \to S$ .
Generalization of Orlov: This generalizes Orlov's Lemma 2.12 (which required smooth projective varieties over a field) to the singular setting and arbitrary base fields.
Methodological Shift: Unlike Orlov's approach, which relied on "canonical morphisms" between kernels describing adjunction units/counits (tools often unavailable in singular settings), this paper uses local algebra and descent to bypass the need for these specific morphisms.

C. Applications to Arithmetic and Geometry

The authors apply these general theorems to derive new invariants for FM partners:

Forms of Curves (Proposition 1.4): For geometrically integral projective curves over a field $k$ of characteristic zero, if they are FM partners, they are $k$ -forms. That is, they become isomorphic after a finite field extension $L/k$ . This extends results from algebraically closed fields to arbitrary fields.
Genus Invariance (Corollary 5.1): For fibrations over a Noetherian $\mathbb{Q}$ -scheme, if the total spaces are FM partners, the genus of the special fibers must coincide.
Smoothness Invariance (Lemma 5.2, Proposition 1.5):
- Smoothness is a derived invariant even over imperfect fields.
- If $Y_1$ and $Y_2$ are FM partners over $S$ , then the morphism $f_1: Y_1 \to S$ is smooth if and only if $f_2: Y_2 \to S$ is smooth.
Cohen-Macaulay and Gorenstein Properties (Proposition 1.6, Proposition 5.3):
- The Cohen-Macaulay and Gorenstein properties of the morphism are preserved under FM partnership for regular Noetherian base schemes.
- Proposition 5.3 strengthens this for Gorenstein properties even when the equivalence does not arise from an integral transform (pure derived equivalence), using dualizing sheaves and intrinsic characterizations of perfect complexes.

4. Significance and Impact

Singular Geometry: The paper successfully bridges the gap between the well-understood smooth case and the difficult singular case, providing robust criteria for derived equivalences without requiring smoothness.
Arithmetic Geometry: By working over arbitrary fields (including imperfect ones) and DVRs, the results are directly applicable to arithmetic problems, such as studying families of varieties over number fields or local rings.
New Techniques: The "yoga" developed—specifically the reliance on local algebra and descent rather than canonical morphisms—offers a new toolkit for researchers studying derived categories in non-smooth or arithmetic contexts.
Future Directions: The authors note that these results are currently being extended to algebraic stacks and moduli spaces, suggesting a broad applicability of their framework.

In summary, this paper provides a comprehensive and rigorous foundation for testing derived equivalences in the most general settings (singular, arithmetic, relative), effectively generalizing the "yoga" of Fourier–Mukai partners to the full scope of modern algebraic geometry.