🔢 mathematics

On the geometry of spaces of filtrations on local rings

This paper investigates the geometry of spaces of saturated filtrations on Noetherian local domains by introducing a Darvas-inspired geodesic metric, characterizing the toric case via Newton-Okounkov bodies and $L^1_\mathrm{loc}$ , analyzing the semi-continuity of log canonical thresholds, and establishing a generalized lattice structure.

Original authors: Lu Qi

Published 2026-04-09

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Lu Qi

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 the shape of a very strange, invisible city. This city isn't made of buildings, but of mathematical rules called "filtrations."

In this paper, the author, Lu Qi, is trying to draw a map of this city. He wants to know: How far apart are two different sets of rules? Can you walk smoothly from one rule-set to another? And what does the "ground" look like underneath?

Here is the story of the paper, broken down into simple concepts.

1. The City of "Filtrations" (The Rules)

First, what is a filtration?
Imagine you have a bag of marbles (the ring $R$ ). A filtration is a way of sorting these marbles into boxes labeled with numbers (like 1.0, 1.1, 1.2...).

The Rule: If a marble is in box 1.5, it must also be in box 1.0, 0.5, and so on. The boxes get smaller as the numbers get bigger.
The Goal: Mathematicians have been studying these boxes for years because they help solve problems in geometry and physics. But until now, nobody had a good way to measure the "distance" between two different ways of sorting these marbles.

2. The New Ruler: The "Darvas Metric"

The author introduces a new ruler, called the $d_1$ metric.

The Analogy: Imagine you have two different ways of organizing a library (two filtrations). To measure how different they are, you look at the books they share versus the books they don't.
The Problem: Sometimes, two very different sorting methods end up looking identical when you only count the total number of books (this is called having the same "multiplicity"). It's like two different recipes that use the exact same amount of flour; they taste different, but the scale says they are the same.
The Fix: The author realizes that to get a true measurement, you have to look at the "saturated" versions of these rules. Think of "saturation" as filling in all the gaps in your sorting method until it's perfectly smooth and complete. Once you do this, the ruler works perfectly.

3. The Smooth Path (Geodesics)

Once the ruler is fixed, the author discovers something amazing: You can walk.

The Analogy: If you want to go from Sorting Method A to Sorting Method B, you don't have to jump instantly. You can take a smooth, continuous path.
The Result: The space of these rules is a geodesic space. This means that between any two points (two different sorting methods), there is a "straight line" (a geodesic) that represents the most efficient way to transform one into the other. It's like finding the shortest flight path between two cities on a globe.

4. The Special Case: The "Toric" City (Lego Blocks)

The paper gets even more concrete when looking at a specific type of city called a Toric Singularity.

The Analogy: Imagine a city built entirely out of Lego blocks arranged in a perfect grid.
The Discovery: In this specific case, the author proves that the entire complex space of sorting rules can be flattened out and mapped onto a simple shape: a convex polygon (like a triangle or a square) with a hole in the middle.
Why it matters: This turns a super-complex, abstract math problem into a simple geometry problem. You can now measure the distance between two sorting methods just by calculating the area of the difference between their shapes. It's like turning a 3D puzzle into a 2D drawing.

5. The "Log Canonical Threshold" (The Sensitivity Gauge)

The paper also studies a specific number called the Log Canonical Threshold (LCT).

The Analogy: Think of the LCT as a "sensitivity gauge" or a "tipping point." It tells you how close a sorting method is to breaking or becoming unstable.
The Finding: The author shows that if you make a tiny change to your sorting method (walking a tiny step on the map), the sensitivity gauge doesn't jump wildly. It changes smoothly. This is crucial because it means the system is stable and predictable. Small errors in measurement won't cause the whole system to collapse.

6. The Lattice Structure (The Lego Tower)

Finally, the author looks at how these rules fit together.

The Analogy: Imagine you have two different Lego towers. You can combine them in two ways:
1. Intersection (Meet): Take only the bricks that are in both towers.
2. Join: Stack them together to make a bigger tower.
The Result: The author proves that these operations work perfectly, creating a structure called a lattice. It's like a perfectly organized filing system where every pair of folders has a clear "smallest common folder" and a "largest combined folder."

Summary: Why Should You Care?

This paper is like building a new GPS for a hidden mathematical world.

It gives us a map: We can now measure distances between abstract mathematical rules.
It gives us roads: We know how to travel smoothly between these rules.
It simplifies the complex: In special cases, it turns high-dimensional math into simple shapes we can draw.
It ensures stability: It proves that small changes in our rules lead to small changes in the results, which is vital for solving real-world problems in physics and engineering that rely on these mathematical models.

In short, Lu Qi has taken a messy, abstract cloud of mathematical possibilities and organized it into a structured, navigable, and stable landscape.

1. Problem Statement and Context

The paper investigates the geometric structure of spaces of filtrations on a Noetherian local domain $(R, \mathfrak{m})$ . Filtrations are defined as continuously indexed families of ideals $\{a_\lambda\}_{\lambda \in \mathbb{R}_{>0}}$ that are decreasing, multiplicative, and left-continuous. They generalize graded sequences of ideals, which have been central to recent developments in algebraic geometry, particularly in the study of $K$ -stability, log canonical thresholds, and singularities.

While the geometry of spaces of norms and filtrations has been extensively studied in the global setting (e.g., on polarized varieties or section rings, as in the work of Darvas, Boucksom, Jonsson, and others), the local setting (on local rings) lacked a unified metric framework. The primary goal of this paper is to construct a metric geometry for the space of filtrations on a local ring, analogous to the Darvas metric in complex geometry and the non-Archimedean metrics developed by Boucksom-Jonsson.

2. Methodology

The author employs a combination of algebraic geometry, valuation theory, and metric space theory. Key methodological components include:

Saturation: A central concept introduced is the saturation of a filtration ( $e a_\bullet$ ). A filtration is saturated if it is determined entirely by its values on divisorial valuations. This resolves the issue where distinct filtrations might have the same multiplicity, allowing for the definition of a true metric on the quotient space.
Multiplicity and Volume: The paper utilizes the multiplicity $e(a_\bullet)$ of a filtration, defined as the asymptotic limit of the colengths $\ell(R/a_m)/m^n$ . This serves as the volume functional analogous to the energy functional in Kähler geometry.
Geodesic Construction: The author defines geodesics between two filtrations $a_{\bullet,0}$ and $a_{\bullet,1}$ via a convolution-like formula:
$a_{\lambda, t} = \sum_{(1-t)\mu + t\nu = \lambda} a_{\mu, 0} \cap a_{\nu, 1}$
This construction is shown to minimize the distance in the proposed metric.
Newton-Okounkov Bodies: In the toric case, the author utilizes Newton-Okounkov bodies to identify the space of saturated monomial filtrations with a space of convex sets, allowing for explicit geometric calculations.
Valuation Spaces: The study relies heavily on the space of valuations centered at the maximal ideal, specifically divisorial valuations ( $DivVal_{R,\mathfrak{m}}$ ) and valuations with positive volume ( $Val^+_{R,\mathfrak{m}}$ ).

3. Key Contributions and Results

A. The Darvas Metric ( $d_1$ ) and Geodesic Structure

The paper introduces a pseudometric $d_1$ on the space of linearly bounded filtrations ( $Fil_{R,\mathfrak{m}}$ ):
$d_1(a_\bullet, b_\bullet) := 2e(a_\bullet \cap b_\bullet) - e(a_\bullet) - e(b_\bullet)$

Main Result (Theorem 1.1): The Hausdorff quotient of $(Fil_{R,\mathfrak{m}}, d_1)$ is the space of saturated filtrations ( $Fils_{R,\mathfrak{m}}$ ).
Geodesic Property: If $R$ contains a field, $(Fils_{R,\mathfrak{m}}, d_1)$ is a geodesic metric space. The geodesic connecting two saturated filtrations is given by the saturation of the segment defined above. This establishes a local analogue of the geodesic metric spaces found in complex geometry.

B. The Toric Case and Isometry to Convex Sets

In the specific case where $(R, \mathfrak{m})$ is the local ring of a toric singularity, the author provides a concrete geometric realization:

Theorem 1.2: The space of saturated monomial filtrations $(Fils^{mon}_{R,\mathfrak{m}}, d_1)$ is isometric to the space of closed convex subsets of the dual cone $\sigma^\vee$ with bounded complements, equipped with the symmetric difference metric (Fréchet-Nikodym-Aronszajn metric).
This identification maps a filtration to its Newton-Okounkov body, proving that the metric space is complete in the toric setting.

C. Supnorm Metric ( $d_\infty$ ) and Homogeneous Norms

The paper introduces a second metric, $d_\infty$ , inspired by the uniform metric on continuous functions:
$d_{\infty, \rho}(\chi, \chi') := \limsup_{\lambda \to \infty} \sup_{\rho(f) \ge \lambda} \frac{|\chi(f) - \chi'(f)|}{\rho(f)}$

Theorem 1.3: Restricted to the space of homogeneous norms (power-multiplicative norms), $d_\infty$ is a metric, and the space is complete.
Comparison: The $d_\infty$ topology is strictly finer than the $d_1$ topology.

D. Semi-continuity of Log Canonical Thresholds

The author studies the continuity of the log canonical threshold (lct) function with respect to these topologies:

Theorem 1.4:
1. Weak Topology: If a sequence of filtrations converges weakly, the lct is lower semi-continuous ( $lct(a_\bullet) \le \liminf lct(a_{\bullet, k})$ ).
2. $+$ -Topology: If a sequence of saturated filtrations converges in the $+$ -topology (convergence of values on all valuations with positive volume), the lct is upper semi-continuous.
3. Lipschitz Continuity: The lct function is locally Lipschitz continuous with respect to the $d_\infty$ metric.

E. Lattice Structure

The paper generalizes the classical result that ideals form a modular lattice:

Theorem 1.5: The space of $m$ -filtrations forms a lattice under inclusion, with meet given by intersection and join defined via a specific construction.
The space of saturated filtrations forms a distributive lattice. The saturation map acts as a surjective join morphism from the lattice of integrally closed ideals to the lattice of saturated filtrations.

4. Significance and Impact

Unification of Local and Global Theories: This work successfully translates the powerful metric geometry of global K-stability (Darvas metrics, geodesics) into the local setting of singularities. This is crucial for the local study of K-stability and the minimization of normalized volumes.
New Tools for Singularity Theory: By establishing that the space of saturated filtrations is a geodesic metric space, the paper opens the door to applying variational methods and geometric analysis techniques to problems involving local singularities, such as the existence of Kähler-Ricci solitons or the study of Kollár valuations.
Structural Insights: The identification of the toric case with convex sets provides a concrete model for understanding the geometry of these infinite-dimensional spaces. The lattice structure results deepen the understanding of the algebraic relationships between ideals and their asymptotic filtrations.
Continuity of Invariants: The semi-continuity results for the log canonical threshold provide robust analytical tools for studying the stability of singularities under deformation or limits of filtrations.

In summary, Lu Qi's paper establishes a rigorous geometric framework for spaces of filtrations on local rings, proving they possess rich metric and lattice structures that generalize classical results and offer new avenues for research in algebraic geometry and singularity theory.