Log-correlated Gaussian Fields Demonstrate Weak Exponential Metrics with Hausdorff Dimension and KPZ Relation

The behaviour of complex systems often hinges on understanding the subtle relationships within seemingly random data, and recent work focuses on quantifying these relationships in high-dimensional spaces. Andres A. Contreras Hip from the University of Chicago and Zijie Zhuang from the University of Pennsylvania, along with their colleagues, demonstrate a rigorous framework for analysing these connections using a novel approach to exponential metrics. Their research establishes that these metrics consistently define meaningful distances within complex, high-dimensional data, extending established theories from two dimensions to more complex spaces. This achievement unlocks new possibilities for understanding phenomena governed by random fields, offering insights into areas ranging from materials science to mathematical physics and providing a foundation for further exploration of complex systems.

The researchers prove that every subsequential limit of exponential metrics, constructed from appropriate approximations of a random field, constitutes a weak γ-exponential metric. Furthermore, the team establishes general properties applicable to any weak exponential metric, specifically sharp moment bounds for several natural distances, and demonstrates optimal Hölder exponents when comparing the field to the Euclidean metric. They also establish results concerning Hausdorff dimension and a KPZ relation, extending the two-dimensional Liouville Quantum Gravity metric theory to higher dimensions and deriving several useful properties for log-correlated Gaussian fields, including the equivalence between different methods of analysis.

Liouville Quantum Gravity and Gaussian Free Fields

This document presents a comprehensive overview of Liouville Quantum Gravity (LQG), Gaussian Free Fields (GFF), and related topics in probability and mathematical physics. LQG seeks to define a natural metric on random surfaces, arising as the scaling limit of discrete random planar maps, and is deeply connected to 2D quantum gravity and string theory. The GFF, a random distribution serving as the mass or potential for defining the LQG metric, is a fundamental object in this study, and understanding its regularity and properties is crucial. The document also references the connection between LQG and discrete random surfaces, where the LQG metric is often defined as the scaling limit of metrics on these discrete surfaces.

LQG surfaces are typically fractal, and the document explores the scaling limits of these fractal objects. The study of LQG relies heavily on tools from probability theory, stochastic analysis, and functional analysis, and has connections to 2D quantum gravity, string theory, and conformal field theory. Key researchers in the field include Robert Adler, Jian Ding, and Julien Dubédat, alongside Ewain Gwynne, Nina Holden, Jason Miller, Grégory Miermont, Scott Sheffield, and Joshua Pfeffer.

Weak Exponential Metrics in Liouville Quantum Gravity

Scientists have established a rigorous mathematical framework for understanding high-dimensional extensions of Liouville Quantum Gravity (LQG). The work centers on log-correlated Gaussian fields and demonstrates the existence of limiting metrics obtained through a process of approximation. Researchers proved that these limiting metrics, termed “weak exponential metrics”, satisfy a defined set of axioms, ensuring their mathematical consistency and well-behaved properties. The team meticulously analyzed the behavior of distances within these weak exponential metrics, establishing sharp bounds on their moments and confirming optimal Hölder exponents when compared to Euclidean distance.

These results extend the two-dimensional LQG metric theory to higher dimensions, opening new avenues for research in areas like random geometry and probability. Experiments involved defining a process of mollification, smoothing the log-correlated Gaussian field using a carefully chosen kernel function. The team investigated the exponential metric derived from this smoothed field, normalizing it with a constant determined by the median distance between two points. Data shows that as the smoothing parameter approaches zero, the family of metrics converges, inducing the Euclidean topology on the space, and confirms the tightness of these metrics. Furthermore, the research establishes a connection between parameters characterizing the log-correlated Gaussian field and the Hausdorff dimension of the weak exponential metric, providing a solid foundation for further exploration of high-dimensional LQG and its applications.

Weak Exponential Metrics in Random Landscapes

This research establishes a rigorous mathematical foundation for understanding distances within complex, random landscapes described by log-correlated Gaussian fields in dimensions higher than two. Scientists have demonstrated the existence of consistent, well-behaved metrics that emerge as one examines these fields at increasingly fine scales. The team proved that subsequences of rescaled exponential metrics converge to what they term ‘weak exponential metrics’, possessing specific properties including predictable bounds on distances between points and sets within the field, extending previous work limited to two dimensions. The researchers further characterized these weak exponential metrics, establishing moment bounds for various types of distances and demonstrating their consistency across different scales. They acknowledge that proving full scaling invariance remains an open challenge and plan to build upon this work to address this in future investigations, intending to explore the uniqueness of these weak exponential metrics and establish full convergence of the rescaled exponential metrics, mirroring results already achieved in two-dimensional cases.