KPZ Relation in Random Geometry

• KPZ relation is a universal quadratic law relating scaling exponents in Euclidean and quantum geometries, providing a bridge between classical and random fractal measures.
• It underpins models in stochastic interface growth by linking roughness, growth, and dynamic exponents, enabling precise predictions in one-dimensional processes.
• Recent rigorous approaches extend the KPZ framework to higher dimensions and dependent fractals using tools like Gaussian multiplicative chaos and advanced diffusion methods.

The KPZ relation describes a universal quadratic law connecting scaling exponents or fractal dimensions in random geometry, stochastic interface growth, and related models of statistical mechanics. Arising originally from the Kardar–Parisi–Zhang (KPZ) universality class for one-dimensional stochastic interface growth, it provides a bridge between exponents in "Euclidean" geometry (classical measure) and "^^^^1^^^^" or random geometries arising in Liouville quantum gravity and related fields. Rigorous formulations and generalizations of the KPZ relation have become central in probability theory, mathematical physics, and geometry.

1. Classical KPZ Relation: Quadratic Exponent Laws

The classical KPZ formula originated in the context of Liouville quantum gravity (LQG) on planar domains, where the geometry is deformed by exponentiating a Gaussian free field (GFF). For $\gamma \in (0,2)$, deterministic fractal sets $A$ have Euclidean (Lebesgue) scaling dimension $x$ related to quantum (LQG) dimension $\Delta$ by: $$x = \frac{\gamma^2}{4}\Delta^2 + \left(1 - \frac{\gamma^2}{4}\right)\Delta$$ or, equivalently,

$$d_{\mathrm{Eucl}} = (2 + \tfrac12 \gamma^2)d_{\mathrm{Q}} - \tfrac12 \gamma^2 d_{\mathrm{Q}}^2$$

This formula describes how the random geometry induced by $e^{\gamma h(z)} dz$, with $h$ a GFF, deforms fractal sets, translating Hausdorff or Minkowski dimensions from Euclidean to LQG measure and vice versa. The KPZ law generalizes to higher even dimensions $\mathbb{R}^{2n}$ with: $$x = \Delta \left(n + \tfrac{\gamma^2}{2}\right) + \tfrac{\gamma^2}{4}\Delta^2$$ as established for $\mathbb{R}^4$ by Chen–Jakobson (Chen et al., 2012).

2. KPZ Relation for Stochastic Interface Growth

In the theory of stochastic interface growth, the KPZ equation is a stochastic PDE modeling the evolution of a one-dimensional height field $h(x,t)$: $$\partial_t h = D\, \partial_x^2 h + \lambda (\partial_x h)^2 + \sqrt{2 D \chi} W(x,t)$$ Here $D$ is the diffusion coefficient, $\lambda$ is the nonlinear growth coefficient, $\chi$ is the static compressibility, and $W(x,t)$ is space-time white noise. The scaling exponents:

• $\alpha$: roughness exponent (height fluctuations $\sim L^\alpha$),
• $\beta$: growth exponent (height growth $\sim t^\beta$),
• $z$: dynamic exponent ($z = \alpha / \beta$), obey the KPZ scaling relation in one dimension: $\alpha + z = 2$ For first-passage percolation (FPP), the fluctuation exponent $\chi$ and wandering exponent $\xi$ satisfy the universal KPZ law: $\chi = 2\xi - 1$ Chatterjee provided a rigorous proof of this relation in FPP under existence assumptions for $\chi$ and $\xi$ (Chatterjee, 2011).

3. Rigorous KPZ Relations for Schramm–Loewner Evolution and LQG

A central breakthrough was the translation of KPZ-type laws into rigorous statements connecting fractal dimensions for SLE curves and their images under LQG measures. In the random geometry induced by LQG, Duplantier–Sheffield, Miller–Sheffield, and related works show that, for Borel sets defined as functions of the SLE path, the almost-sure KPZ relation holds: $$\dim_H(X) = (2 + \tfrac{\gamma^2}{2})\, d - \tfrac{\gamma^2}{2}\, d^2$$ This holds for any set $X$ in the range of a space-filling SLE$_{\kappa}$ curve ($\gamma=4/\sqrt{\kappa}$, $\kappa > 4$), when $X$ is a deterministic function of the curve (modulo parametrization). The peanosphere construction reduces the problem to Brownian motion, whose fractal dimensions are classical and well understood, enabling comprehensive calculation of SLE Hausdorff dimensions (Gwynne et al., 2015).

4. Breakdown and Deformations of the KPZ Relation

While the KPZ relation is universal for deterministic and GFF-independent sets, it can fail systematically for natural fractals that depend nontrivially on the underlying Gaussian free field. Aru demonstrated that for the zero-level lines (SLE$_4$) and SLE$_\kappa$ flow lines of a GFF, the expected quantum Minkowski dimensions are strictly less than the KPZ-predicted values. The correct relation for SLE$_\kappa$ flow lines, in the coupling with the GFF via imaginary geometry, is a "deformed" quadratic law: $$d_M = (2 + \tfrac{\gamma^2}{2})\, q_{M,E} - \tfrac{\gamma^2}{2} \left(1 - \frac{\kappa}{4}\right)^2 q_{M,E}^2$$ where $d_M$ is the Euclidean Minkowski dimension and $q_{M,E}$ is the quantum dimension (Aru, 2013). This formula only matches the classical KPZ law in the limits $\kappa \to 0$ or $\kappa \to 8$. The key technical mechanism is the introduction of winding-dependent correction terms due to nontrivial coupling between geometry and measure.

Curve Type / Setting	KPZ-type Relation	Regime/Restriction
Planar LQG, GFF-independent set	$$d_{\mathrm{Eucl}} = (2+\frac12 \gamma^2) d_{\mathrm{Q}} - \frac12 \gamma^2 d_{\mathrm{Q}}^2$$	Universal for $0<\gamma<2$
SLE$_\kappa$ flow line, GFF-coupled	$$d_M = (2+\frac12 \gamma^2) q_{M,E} - \frac12 \gamma^2 (1-\frac{\kappa}{4})^2 q_{M,E}^2$$	Imaginary geometry, $0<\kappa<8$
First-passage percolation exponents	$\chi = 2\xi - 1$	Subadditive ergodic model
4D LQG (Chen–Jakobson)	$$x = \Delta(2+\frac{\gamma^2}{2}) + \frac{\gamma^2}{4}\Delta^2$$	$0<\gamma^2<2\pi^2$

5. KPZ-Type Relations in Other Contexts

First-Passage Percolation

The KPZ scaling relation $\chi = 2\xi - 1$ connects fluctuation and wandering exponents in FPP, with rigorous proofs relying on subadditive ergodic theory, curvature estimates of the limit shape, and thin-cylinder variance decomposition methods (Chatterjee, 2011). This relation is conjectured to hold universally across dimensions and is independent of the lattice structure or edge distribution, conditional on existence of the exponents.

Higher-Dimensional and Manifold Generalizations

The KPZ paradigm extends to higher even dimensions. For the canonical GFF on $\mathbb{R}^4$ with correlation kernel $G(\epsilon) \sim -\frac{1}{2\pi^2} \log \epsilon$, the KPZ law reads: $$x = \Delta(2 + \tfrac{\gamma^2}{2}) + \frac{\gamma^2}{4}\Delta^2$$ as proved by Chen–Jakobson, with analogous expressions proposed for $\mathbb{R}^{2n}$ and conformally invariant settings on 2n-dimensional manifolds (Chen et al., 2012).

6. Key Analytical Inputs and Methodologies

• Gaussian Multiplicative Chaos is the standard technique to construct the random measure $e^{\gamma h(z)} dz$, where $h$ is a GFF, requiring renormalization and circle-average approximations.
• Martingale Methods underpin both LQG measure construction and computations of moment scaling, often reduced to exponential martingales and Feynman–Kac-type formulae.
• Brownian and Diffusion Methods: SLE dimension results and KPZ reduction can be formulated probabilistically via Brownian motion, special diffusions (e.g., winding-number processes), or the peanosphere encoding.
• Deformation Mechanisms: For GFF-dependent SLE, winding correction factors appear in the moment generating function, leading to altered KPZ exponents as established by precise diffusion analysis and quantum ball covering estimates (Aru, 2013).

7. Open Problems, Generalizations, and Programmatic Directions

• Determining necessary and sufficient conditions for the validity of KPZ relations for arbitrary GFF-dependent fractals remains open. Only extremal upper and lower bounds are currently conjectured for generic dependencies.
• Systematic study of "deformed" KPZ relations associated with other SLE-type objects (CLE$_\kappa$, SLE$_{\kappa,\rho}$), multiple-arm events, and their critical exponents is ongoing.
• Higher even-dimensional and manifold extensions, utilizing critical GJMS operators and generalizing LQG, continue to be developed, but rigorous results are established only in flat space for $n=2,4$.
• Joint winding estimates of SLE-driven sets conditioned on passing near multiple points, which may enable almost-sure quantum dimension estimates for coupled curves, are an outstanding technical challenge.

KPZ relations thus provide a unifying framework for scaling in random geometries, but their validity and form crucially depend on the independence structure between sets and underlying fields. Contemporary work rigorously establishes both universality and systematic failure of the classical KPZ law in dependent regimes, with progress driven by analytical control of multiplicative chaos and detailed study of fractal and winding structures in planar stochastic models (Aru, 2013, Chen et al., 2012, Gwynne et al., 2015, Chatterjee, 2011).