Analysis of PDEs
Papers
Fresh from arXiv, typeset properly. Highlighted words trade on the market.
Tuesday 12 May
This paper presents a novel time-space SIR (Susceptible-Infected-Recovered) model for simulating infectious disease dynamics in two interconnected regions. The model is formulated as a coupled reaction-diffusion system with boundary conditions that dynamically switch from Robin to Neumann types, effectively modelling policy-driven interventions such as lockdowns. A key innovation lies in the incorporation of degenerate diffusion, arising from vanishing population density, which significantly influences transmission behaviour near regional borders. The wellposedness of the model is rigorously established using the Faedo-Galerkin method, ensuring the existence, uniqueness, and positivity of weak solutions. Numerical simulations, performed using the Finite Volume Method, validate the theoretical findings and demonstrate the impact of migration and mobility restrictions on epidemic progression. This framework offers valuable insights for understanding and controlling disease spread in spatially heterogeneous and interconnected settings.
Information Theory
Constacyclic codes of length $np^s$ over $\frac{\mathbb{F}_{p^m}[u]}{\langle u^t\rangle}$: Torsions and Cardinalities
The purpose of this article is to study constacyclic codes of length $np^s$ over $R^t:=\frac{\mathbb{F}_{p^m}[u]}{\langle u^t \rangle },$ where $t$ is a natural number and $\gcd(n,p)=1$. We give generators of all the ideals of $R^{t,n}_{\delta}:=\frac{R^t[x]}{\langle x^{np^s}-\delta \rangle},$ where $\delta= \delta_0+u\delta_1+\dots+u^{t-1}\delta_{t-1}$ is a unit in $R^t$. For $n=1,\ 2, \ 3$ and $t=3$, we provide all types of ideals (constacyclic codes) and also give the torsional degrees as well as cardinalities of these codes.
Read the full abstract
Algebraic Geometry
Positive Geometries from Cubic Surfaces
We present a study of cubic surfaces from the novel perspective of positive geometry. Our positive geometries have dimension two (the surface minus its 27 lines), dimension three (its complement in 3-space), and dimension four (the moduli space). In each case we explore the positive arrangement, its combinatorial rank, and the canonical forms.
Read the full abstract
Representation Theory
On quiver skew braces, their ideals and products
Quiver skew braces or skew bracoids are equivalent to braided groupoids, that is, groupoids with a constraint of abelianity. They are the quiver-theoretic version of skew braces, an increasingly studied structure lying in the intersection of group and ring theory. In this paper, we define ideals and quotients for quiver skew braces, with respect to two notions of morphisms. Following the track of a previous work of ours (2025), we define a classical semidirect product à la Brown, and a categorical semidirect product à la Bourn and Janelidze, for the category of quiver skew braces. It is known that connected groupoids can be expressed as the datum of a group of loops and a set of vertices. We demonstrate how no such decomposition holds for quiver skew braces, which makes their theory richer than the theory of groupoids.
Read the full abstract
Numerical Analysis · Cross-listed from physics.flu-dyn
A Volume of Fluid Immersed Boundary Method for Industrial Polymer Mixing
This work develops advanced numerical methods for free-surface simulations of polymer mixing processes, integrating a Volume of Fluid (VOF) interface-capturing approach with a non-conforming Immersed Boundary (IB) method to model two-phase flows of highly viscous polymer melts and air within partially filled rotating mixing devices, implemented within the Finite Volume OpenFOAM library. To overcome severe numerical instabilities arising from the strong viscosity contrast between polymer melts and air, a block-coupled scheme providing fully implicit viscous diffusion treatment is integrated into the VOF-IB framework, relaxing time-step stability constraints and substantially reducing computational cost with respect to standard segregated solvers. The resulting BC-VOF-IB solver is applied to industrially relevant geometries of single- and twin-screw extruders, yielding physically consistent predictions of velocity and pressure fields under partial filling conditions. While further developments, most notably the inclusion of thermal effects, remain necessary, the proposed framework represents a meaningful step toward bridging academic CFD research and the practical demands of industrial polymer processing.
Read the full abstract
Number Theory
Bigness of Canonical Quadratic Points on Curves of Genus 4
A central problem in arithmetic geometry is to construct non-torsion rational points on elliptic curves. We study a canonical quadratic point $\xi_C \in {\rm Jac}(C)$ attached to a smooth non-hyperelliptic curve of genus 4 and use it to produce such points on elliptic curves arising from families of genus $4$ curves. We introduce a notion of bigness for sections of abelian schemes and establish a criterion in terms of modular variation of abelian quotients, using adelic line bundles and Betti maps. As applications, we prove that $\xi_C$ is big on the triple-involution locus and on certain CM families, obtaining in particular non-torsion rational points on the associated elliptic curves and Northcott-type finiteness results.
Read the full abstract
Functional Analysis
A comparison of the weakest contractive conditions for Banach and Kannan mappings
We study the weakest convergence-type conditions for fixed point results for Banach and Kannan mappings. Building on Suzuki's weakest condition for Banach mappings and our previous result for Kannan mappings, we compare convergence conditions defined along Picard sequences. We give a direct proof that several weakest convergence conditions are equivalent for Kannan-type mappings on complete metric spaces. This proof is achieved without assuming the completeness or the convergence of Picard sequences; it deduces the equivalence only from the existence of fixed points. In contrast, we construct a counterexample showing that the corresponding equivalence fails for Banach contractions. Finally, we prove that this discrepancy disappears on G-complete metric spaces, clarifying the role of completeness in weakest fixed point theory.
Read the full abstract
Probability
Fluctuation analysis for a randomly perturbed dynamical system with state-dependent impulse effects
The principal aim of the present work is to explore limit theorems for small random perturbations of a planar impulsive dynamical system, where impulses occur at hitting times of a suitable switching surface, and are thus state-dependent. Working with a simplified example in polar coordinates, we obtain-for any fixed time horizon-a small noise expansion for the radial component, together with rigorous error estimates in the Skorohod space of right-continuous functions with left limits.
Read the full abstract
Information Theory · Cross-listed from cs.LG
Information theoretic underpinning of self-supervised learning by clustering
Self-supervised learning (SSL) is recognized as an essential tool for building foundation models for Artificial Intelligence applications. The advances in SSL have been made thanks to vigorous arguments about the principles of SSL and through extensive empirical research. The aim of this paper is to contribute to the development of the underpinning theory of SSL, focusing on the deep clustering approach. By analogy to supervised learning, we formulate SSL as K-L divergence optimization. The mode collapse is prevented by imposing an optimisation constraint on the teacher distribution. This leads to normalization using inverse cluster priors. We show that using Jensen inequality this normalization simplifies to the popular batch centering procedure. Distillation and centering are common {heuristics-based} practices in SSL, {but our work underpins them theoretically.} The theoretical model developed not only supports specific existing successful SSL methods, but also suggests directions for future investigations.
Read the full abstract
Optimization and Control
Constrained Stochastic Spectral Preconditioning Converges for Nonconvex Objectives
In this work, we develop proximal preconditioned gradient methods with a focus on spectral gradient methods providing a proximal extension to the Muon and Scion optimizers. We introduce a family of stochastic algorithms that can handle a wide variety of convex and nonconvex constraints and study its convergence under heavy-tailed noise, through a novel analysis tailored to the geometry of the proposed methods. We further propose a variance-reduced version, which achieves faster convergence under standard noise assumptions. Finally, we show that the polynomial iterations used in Muon are more accurately captured by a nonlinear preconditioner than by the ideal matrix sign, leading to a convergence analysis that more faithfully reflects practical implementations.
Read the full abstract
Mathematical Physics · Cross-listed from nlin.CD
Stochastically perturbed billiards: fingerprints of chaos and universality classes
Billiards tables - a minimal model for particles moving in a confined region - are known to present classical (and quantum) different features according to their shape, ranging from strongly chaotic to integrable dynamics. Here we consider the role of a stochastic perturbation of the elastic reflection law, and show that while chaotic billiards maintain their key statistical feature, the behaviour for integrable billiard tables is completely different: it can be linked, for tiny perturbations, to Evans stochastic billiard, where at each collision the reflected angle is a uniformly distributed stochastic variable on $(-\pi/2,\pi/2$). The resulting spatial stationary measure has peculiar aspects, like being typically non uniform along the boundary, differently from any chaotic billiard table.
Read the full abstract
Dynamical Systems
Cocycles with Quasi-Conformality II: Ergodic measures with positive entropy
As the second part of a series on linear cocycles over chaotic systems, this paper establishes a "multiple covering principle" that robustly yields positive-entropy ergodic measures supported on fiberwise uniformly bounded orbits. Using this mechanism, we prove that any continuous $\mathrm{SL}(d,\mathbb{R})$ cocycle over a positive-entropy subshift of finite type either admits a dominated splitting or can be $\mathcal{C}^0$-approximated by one that stably supports such measures. The stability holds in Hölder topology which is optimal. Additionally, for non-isometric cocycles, we show that the topological entropy of these bounded orbits is strictly less than that of the base subshift.
Read the full abstract
Numerical Analysis
Dynamics of the Longest-Edge Altitude Bisection Algorithm
We study a longest-edge based refinement scheme for triangulations, termed the longest-edge altitude bisection (LEAB), in which each triangle is subdivided by dropping the altitude from the vertex opposite to its longest edge. Using the normalized shape space of triangles introduced by Perdomo and Plaza in: Properties of triangulations obtained by the longest-edge bisection. Cent. Eur. J. Math., 12(12) (2014), 1796-1810, we show that LEAB admits a simple geometric description: the normalized left and right children of a triangle in focus are obtained by intersecting the geodesic of right triangles with rays issued from the endpoints of the longest edge and explicit formulas for the mappings are derived. This characterization implies an interesting observation that the associated refinement dynamics collapse the entire shape space onto the right-triangle geodesic in a single step and that every point on this geodesic is fixed. Two-sided bounds for the contraction of the mesh size (discretization parameter) are derived. Also, applications and limitations of the method are briefly discussed.
Read the full abstract
Optimization and Control · Cross-listed from cs.LG
Gradient Clipping Beyond Vector Norms: A Spectral Approach for Matrix-Valued Parameters
Gradient clipping is a standard safeguard for training neural networks under noisy, heavy-tailed stochastic gradients; yet, most clipping rules treat all parameters as vectors and ignore the matrix structure of modern architectures. We show empirically that data outliers often amplify only a small number of leading singular values in layer-wise gradient matrices, while the rest of the spectrum remains largely unchanged. Motivated by this phenomenon, we propose spectral clipping, which stabilizes training by clamping singular values that exceed a threshold while preserving the singular directions. This framework generalizes classical gradient norm clipping and can be easily integrated into existing optimizers. We provide a convergence analysis for non-convex optimization with spectrally clipped SGD, yielding the optimal $\mathcal{O}\left(K^{\frac{2 - 2\alpha}{3\alpha - 2}}\right)$ rate for heavy-tailed noise. To minimize hyperparameter tuning, we introduce layer-wise adaptive thresholds based on moving averages or sliding-window quantiles of the top singular values. Finally, we develop efficient implementations that clip only the top $r$ singular values via randomized truncated SVD, avoiding full decompositions for large layers. We demonstrate competitive performance across synthetic heavy-tailed settings and neural network training tasks.
Read the full abstract
Analysis of PDEs
Sharp upper bound for a branched transport problem coming from Ginzburg-Landau models
We consider a branched transport type problem with weakly imposed boundary conditions, which can be seen as a blown-up version of a reduced model for type-I superconductors in the regime of vanishing external magnetic field. We prove that if the irrigated measure is (locally) Ahlfors regular then it is of dimension at most $8/5$ in agreement with the conjecture by Conti, the third author and Serfaty.
Read the full abstract
Metric Geometry
Self-similar dendrites with finite boundary and P-sprouts
Each self-similar dendrite K with a finite self-similar boundary defines a finite acyclic edge-labeled bipartite graph G, called the sprout of K. The paper shows that the sprout G determines the combinatorial properties of the dendrite K and its topological structure.
Information Theory
Maximum Entropy of Sums of Independent Ternary Random Variables
The classical problem of maximizing the Shannon entropy of a sum of independent random variables supported on a finite alphabet is considered and settled in the ternary case. Namely, the following theorem is established: if \(X_1,\ldots,X_n\) are independent random variables taking values in \(\{0,1,2\}\), then the entropy of \(S_n=X_1+\cdots+X_n\) is maximized when \(X_1,\ldots,X_{n-1}\) are uniform on \(\{0,2\}\) and the probability mass function of \(X_n\) is given by \(\Prob(X_n=0) = \Prob(X_n=2) = w/2\), \(\Prob(X_n=1) = 1-w\), where \(w = \big(1 + 2^{-H(B_n)+H(B_{n-1})}\big)^{-1}\) and \(B_m\sim \Bin(m,1/2)\). The statement can be seen as an extension to ternary alphabets of the Shepp–Olkin–Mateev theorem. The proof uses the Hermite–Biehler theorem, Newton's inequalities, and Yu's maximum-entropy theorem for ultra-log-concave distributions.
Read the full abstract
Information Theory
Polar Complexity: A New Descriptive Complexity with Applications to Source and Joint Source-Channel Coding
This paper first presents a new approach to evaluating the descriptive complexity of finite-length binary sequences. Specifically, we investigate the sequence-wise recovery behavior induced by polar compression and successive cancellation decoding (SCD), and define the polar complexity of a sequence as the minimum polar-compression length (PCL) required for its exact reconstruction. To compute the polar complexity efficiently, we further develop both a bisection-search algorithm and a low-complexity estimation method. We then propose a polar-based two-stage source coding scheme, in which each source sequence is represented by its polar complexity followed by the corresponding polar-compressed sequence. The proposed scheme is strictly lossless and prefix-free. In addition, for BMSs, the normalized average compression length of the proposed scheme can asymptotically approach the source entropy under certain conditions. Simulation results further demonstrate that the scheme can operate without prior knowledge of the source statistics and remains robust across different source distributions. Finally, we integrate the proposed polar source coding with polar channel coding to develop an adaptive double-polar joint source-channel coding (JSCC) scheme, where the encoder and decoder share a predefined set of candidate PCLs to balance error performance and decoding complexity. We formulate the design of the candidate-PCL set as an optimization problem and solve it efficiently via dynamic programming. Simulation results show that the proposed adaptive double-polar JSCC scheme provides a flexible performance-complexity tradeoff and outperforms existing polar-code-based JSCC baselines.
Read the full abstract
Combinatorics
Classification and counting of Gorenstein simplices with $h^*$-polynomial $1+t^k+\cdots+t^{(v-1)k}$
Hibi, Yoshida, and the author classified Gorenstein simplices which are not lattice pyramids and whose \(h^*\)-polynomials are of the form \(1+t^k+t^{2k}+\cdots+t^{(v-1)k}\) when \(v\) is a prime number or the product of two prime numbers. They also conjectured that, for general \(v\), the number of unimodular equivalence classes of such simplices depends only on the divisor lattice of \(v\). This paper proves the conjecture by giving a constructive classification of Gorenstein simplices whose \(h^*\)-polynomials are of this form. More precisely, their unimodular equivalence classes are shown to be parametrized by strict divisor chains in the divisor lattice of \(v\) together with certain recursive combinatorial data. As a consequence, an explicit formula for the number of equivalence classes in terms of the divisor lattice of \(v\) is obtained.
Read the full abstract
Combinatorics
An algebraic-combinatorial framework for finding the average hitting times in graphs with high regularity
For any given vertices $u$ and $v$ in a graph, the hitting time of a random walk on a finite graph is the number of steps it takes for a random walk to reach vertex $v$ starting at vertex $u$. The expected value of the hitting time is the average hitting time. In this paper, we present an algebraic-combinatorial method for calculating the average hitting time between vertices of finite graphs exhibiting high regularity, along with its applications to multiple graph classes. Our approach exploits a novel connection between maximal-entropy random walks and weight-equitable partitions, providing a unifying framework that strengthens and extends several known results, including Rao's method [Statistics & Probability Letters, 2013] for computing the hitting time from a vertex to a neighbor under certain symmetries of the starting vertex.
Read the full abstract
Information Theory
Empirical coordination in the finite blocklength regime: an achievability result—Extended version
Empirical coordination offers a way to understand how agents can coordinate actions under communication constraints. This paper investigates the finite blocklength regime of this problem, where the encoder and decoder aim to produce a sequence of action pairs that is jointly typical with respect to a target distribution. Adopting Shannon's random coding argument and leveraging the method of types, we analyze the average performance of a random codebook to establish an achievability result. The resulting bound on the optimal rate is presented both in exact form and as an asymptotic expansion, aligning with the prevailing characterizations in the finite blocklength literature. This work extends finite blocklength analysis to the empirical coordination setting, complementing existing results on strong coordination.
Read the full abstract
Statistics Theory
Adaptive Kernel Ridge Regression with Linear Structure: Sharp Oracle Inequalities and Minimax Optimality
Kernel ridge regression (KRR) is a widely used nonparametric method due to its strong theoretical guarantees and computational convenience. However, standard KRR does not distinguish between linear and nonlinear components in the signal, instead applying a single functional regularization to the entire function. This may lead to unnecessary shrinkage of linear structure and consequently suboptimal prediction performance. In this paper, we propose a modified regression procedure that augments KRR with an explicit linear component. The proposed method has the same computational complexity as standard KRR and introduces no additional tuning parameters. Theoretically, we establish a sharp oracle inequality for the proposed estimator and show that it adaptively captures both linear and nonlinear structure, achieving minimax optimal prediction risk under general kernels. Compared with standard KRR, the proposed method improves both the bias and approximation error at the expense of only an additional parametric variance term, which is negligible in low- and moderate-dimensional settings. In high-dimensional regimes, incorporating ridge regularization for the linear component yields a procedure that performs uniformly no worse than KRR. Extensive simulation studies support the theoretical findings.
Read the full abstract
Probability
A non-local singular non-linear Fokker-Planck PDE
The focus of this paper is a non-local singular non-linear Fokker-Planck partial differential equation (PDE). The peculiarity of this PDE feature is in its divergence coefficient, which presents a product between a Besov distribution and a non-linearity. The latter involves the convolution between an integrable kernel K and the solution of the PDE, which leads to a non-locality of the first order term in the PDE. We prove existence and uniqueness of a solution to the PDE as well as continuity results on its coefficients. Previous analytical results are then applied to the study of well-posedness in law for a non-local singular McKean stochastic differential equation. As byproduct of that probabilistic representation, we establish mass conservation and positivity preserving for the PDE.
Read the full abstract
Complex Variables
Normality Criteria for Quasimeromorphic Mappings: Avoidance and Sharing
We establish normality criteria for families of quasimeromorphic mappings in \(\mathbb{R}^n\), \(n\geq 2\), based on the avoidance and sharing of continuous functions. We first establish a rescaling principle for sequences of continuous functions and use it to derive a higher-dimensional analogue of Lappan's avoidance criterion. We then obtain normality criteria based on the sharing of continuous functions. In particular, we prove that a family of \(K\)-quasimeromorphic mappings sharing \(q_0(n,K)+1\) mutually disjoint continuous functions is normal, where \(q_0(n,K)\) is the corresponding quasiregular Picard constant. We further consider sequences of quasimeromorphic mappings sharing sequences of continuous functions with a normal comparison sequence and establish a corresponding normality criterion, together with an improved bound under an additional separation condition on the limiting functions. Finally, we derive applications to normal quasiregular mappings into \(\mathbb{S}^n\).
Read the full abstract
Analysis of PDEs
The unified transform for Burgers' equation: Application to unsaturated flow in finite interval
In this paper, we focus on one-dimensional vertical infiltration, assuming constant diffusivity and a quadratic relationship between hydraulic conductivity and water content. Under these assumptions, Richards' equation reduces to Burgers' equation, which we then linearize via the Hopf-Cole transformation. This turns the initial boundary value problem into a diffusion equation on a finite interval with mixed boundary conditions. To solve it, we use the Unified Transform Method (also known as the Fokas method). This approach gives an explicit integral representation of the solution, and when evaluated numerically, the results match classical Fourier series solutions exactly, but with better convergence and stability. Two examples from hydrological applications are examined.
Read the full abstract
Probability
Regularization of a mean-field SDE by an additive common noise: The conditional expectation case
We investigate a McKean-Vlasov stochastic differential equation with an additive common noise and in which the interaction is through the conditional expectation. We show that, in the presence of an additive individual noise, existence and uniqueness of a weak solution hold for any drift given by a bounded and measurable function of the position and the conditional expectation. When there is no individual noise, existence and uniqueness still hold if the drift is in addition Lipschitz in the position variable. This shows that the presence of a finite dimensional common noise may allow to overcome the discontinuity of the drift with respect to the interaction term, provided that this interaction term is a conditional expectation. We also prove propagation of chaos for systems of particles where the conditional expectation is replaced by the empirical mean of the positions or by a closely related contribution with better prepared noise.
Read the full abstract
Probability
Stability of compensator integrals and nonlinear operators under canonical p-th power variation convergence
Let \(X\) and \((X^n)_{n\ge1}\) be c`adl`ag processes with predictable compensators \(\nu\) and \(\nu^n\). For \(p>1\), assume that \([X^n-X]^{(p)}_t\to0\) in probability, where \([\,\cdot\,]^{(p)}\) denotes canonical \(p\)-th power variation. We prove canonical \(p\)-th power-variation stability for a class of nonlinear compensator operators. Their atom responses combine a stable global Lipschitz cylinder component generated by finitely many compensator integrals with a nonlinear residual on a compact part of the mark space. A stopped compensation estimate and weighted cylinder approximation yield stability without assuming topological convergence of \(\nu^n\) to \(\nu\). Examples include covariance, cumulant, kernel and regularised transport responses. As a linear foundation, we prove stability of compensator integrals under local growth, locally uniform convergence of the integrands, jump continuity and barycentric tail control. The proof rests on two structural consequences of canonical \(p\)-th power-variation convergence: threshold isolation, which produces spatial bands eventually avoided by the limiting and approximating jump magnitudes, and compensator mass control, which uses predictable jump matching and localisation to construct an almost surely summable envelope for annular atom masses. These consequences require only the standing process and finiteness assumptions, independently of the additional hypotheses for compensator-integral stability. We also establish sufficient conditions for ucp stability of compensator integrals and, for \(p=2\), an equivalence criterion for ucp stability of the corresponding compensated jump integrals.
Read the full abstract
Representation Theory
Representations of Hecke-Clifford superalgebras at roots of unity
In this article, we give a classification of irreducible completely splittable representations of affine Hecke-Clifford superalgebras $H_n^{\mathrm{aff}}(q)$ when $q^2$ is a primitive $h$-th root of unity. As an application, we derive a necessary and sufficient condition for the finite Hecke-Clifford superalgebra $H_n(q)$ to be semisimple. Specially we show that $H_n(q)$ is semisimple if and only $h >n$ in the case $h$ is odd and $h >2n$ in the case $h$ is even.
Read the full abstract
Differential Geometry
Ricci Curvature and Betti Numbers of Hessian Manifolds
We study Ricci curvature properties of Hessian metrics on the leaves of the codimension-one foliation $\mathcal{F}_\omega = \ker\,\omega$ generated by the first Koszul form $\omega$ of a closed oriented Hessian manifold. Our main result reveals a striking rigidity phenomenon: non-negative Ricci curvature on a single leaf of $\mathcal{F}_\omega$ compels the Hessian metric to be flat, yields sharp bounds on the first Betti number in terms of the dimension of the Hessian manifold and the topology of the leaves. This rigidity also shows that Koszul-type and radiant affine manifolds admit no leaf carrying non-negative Ricci curvature, reflecting a fundamental incompatibility between affine hyperbolicity and leafwise curvature positivity. In dimension three, we obtain a complete classification of the underlying manifold, extended to the non-orientable setting via the orientation double cover.
Read the full abstract
Numerical Analysis · Cross-listed from cs.CE
A nonlinear extension of parametric model embedding for dimensionality reduction in parametric shape design
Dimensionality reduction is essential in simulation-based shape design, where high-dimensional parameterizations hinder optimization, surrogate modeling, and systematic design-space exploration. Parametric Model Embedding (PME) addresses this issue by constructing reduced variables from geometric information while preserving an explicit backmapping to the original design parameters. However, PME is intrinsically linear and may become inefficient when the sampled design space is governed by nonlinear geometric variability. This paper introduces a nonlinear extension of PME, denoted NLPME. The proposed framework preserves the defining principle of PME – geometry-driven latent variables and parameter-mediated reconstruction – while replacing the linear reduced subspace with a nonlinear latent representation. Geometry is not reconstructed directly from the latent variables; instead, the latent representation is decoded into admissible design parameters, and the corresponding geometry is recovered through a forward parametric map. The method is assessed on a bio-inspired autonomous underwater glider with a 32-dimensional parametric shape description and a CAD-based geometry-generation process. NLPME reaches a 5% reconstruction-error threshold with \(N=5\) latent variables, compared with \(N=8\) for linear PME, and a 1% threshold with \(N=9\), compared with \(N=15\) for PME. Comparison with a deep autoencoder shows that most of the nonlinear compression gain can be retained while preserving an explicit backmapping to the original design variables. The results establish NLPME as a compact, admissible, and engineering-compatible nonlinear reduced representation for parametric shape design spaces.