Peer Revenue Past citations are no guarantee of future returns

Papers

Fresh from arXiv, typeset properly. Highlighted words trade on the market.

Mathematics

math Watch Mathematics (sign in to use)

Monday 18 May

Probability

Large deviations of crowding in finite $\beta$-ensembles

We consider finite $\beta$-ensembles $\mathcal X_{n,\beta}^{\mathbb F}$ with $n$ points on $\mathbb F$, where $\mathbb F$ denotes either the real line or the complex plane. Let $U$ be a bounded subset of $ \mathbb F$ such that $\partial U$ (the boundary of $U$) is polar for $\mathbb F=\mathbb R$ and $\partial U$ is a closed $1$–rectifiable set with finite $1$-dimensional Hausdorff measure. Suppose $\mathcal X_{n,\beta}^{\mathbb F}(U)$ denotes the number of points in the region $U$. We show that the sequence of laws of $\{n^{-1}\mathcal X_{n,\beta}^{\mathbb F}(U); n\ge 1\}$ satisfies the large deviation type bound with speed $n^2$ and with a good rate function. For $\mathbb{F} = \mathbb{R}$, this result can be derived using the contraction principle. However, when $\mathbb{F} = \mathbb{C}$, the contraction principle does not yield the desired outcome. Therefore, we adopt a direct approach to establish our results.
Read the full abstract

Functional Analysis

On Multiplicity of Uniform Norms and Maximal Spectral Substructures in Commutative Banach Algebras

Let $\mathcal A$ be a semisimple commutative Banach algebra. It is shown that either $\mathcal A$ has exactly one uniform norm or it admits uncountably many uniform norms. Further, it is shown that there always exists a largest closed subalgebra of $\mathcal A$ which is weakly regular, and that there always exist largest closed ideals in $\mathcal A$ having unique uniform norm property (UUNP) and spectral extension property (SEP) respectively.
Read the full abstract

Optimization and Control · Cross-listed from cs.LG

Ringmaster LMO: Asynchronous Linear Minimization Oracle Momentum Method

Muon has recently emerged as a strong alternative to AdamW for training neural networks, with encouraging large-scale pretraining results and growing evidence that matrix-structured updates can be faster in practice. Yet Muon, and more generally Linear Minimization Oracle (LMO) based methods, are typically used synchronously. This is problematic in heterogeneous distributed systems, where workers complete gradient computations at different speeds and synchronous training must repeatedly wait for slower workers. In this work, we introduce Ringmaster LMO, an asynchronous LMO-based momentum method for unconstrained stochastic nonconvex optimization. Our method builds on the delay-thresholding idea of Ringmaster ASGD. For SGD-type methods, Ringmaster ASGD achieves optimal time complexity by discarding overly stale gradients. Ringmaster LMO extends this mechanism to general LMO-based updates. We establish convergence guarantees under generalized $(L_0, L_1)$-smoothness and further develop a parameter-agnostic variant with decreasing stepsizes and adaptive delay thresholds. Finally, we translate our iteration guarantees into time complexity bounds under heterogeneous worker computation times. In the classical Euclidean smooth setting, these bounds recover the optimal time complexity of Ringmaster ASGD. Experiments on stochastic quadratic problems and NanoChat language-model pretraining show that the advantages of Ringmaster LMO grow with system heterogeneity and that the method outperforms strong synchronous and asynchronous baselines.
Read the full abstract

All authors Abdurakhmon Sadiev, Artavazd Maranjyan, Ivan Ilin, Peter Richtárik

Dynamical Systems

Explicit entropy bounds for symmetric nearest-neighbor subshifts

We provide another approach to Friedland's result that the topological entropy $h$ of a symmetric nearest-neighbor subshift is computable. Instead of the previous algebraic technique, our approach is mostly combinatorial and involves only counts of locally admissible patterns $C_n$ of a cube $[1,n]^d$ in $\mathbb Z^d$. The main idea is a reflection-gluing construction: we flip admissible patterns and merge them along their boundaries. In addition to a short and elementary proof, another advantage is that our approach yields an explicit convergence rate in arbitrary dimensions, whereas obtaining such a rate is already complicated for $\mathbb Z^3$ in Friedland's approach. In particular, we show that for every $n\ge 1$, \[ \frac{1}{n^d}(\log C_{n+1} - q_d(n)\log|\Sigma|) \le h \le \frac{1}{n^d} \log C_n, \] where $\Sigma$ is the alphabet and \[ q_d(n)=(2^d-1)\sum_{k=0}^{d-1} \frac{\binom{d}{k}}{2^d-2^k}\, n^k. \]
Read the full abstract

Optimization and Control

Resolvent Moreau identities without monotonicity: theory and applications to Gabay duality, Douglas–Rachford and ADMM

Duality is most often defined as a relationship between convex functions. If those functions are nonconvex, classical duality breaks down. Notwithstanding, we show that another kind of duality still exists, not between the functions themselves, but between the so-called resolvent operators used to solve associated problems. In fact, this duality-like relationship holds for any set-valued mapping, and is a generalization of the Moreau's identity. We use this duality to study existing operator schemes and to design new ones. In particular, we show that the duality-like relationship Daniel Gabay illuminated between the Douglas–Rachford splitting (DR) and the Alternating Direction Method of Multipliers (ADMM) extends to nonmonotone inclusion problems. We use this relationship to provide explicit counterexamples to the convergence of ADMM in several open cases, by studying the (easier to analyse) DR scheme. Motivated by our observations, we design a class of convergent resolvent homotopy schemes and use them to solve nonconvex-regularised least absolute deviations problems. This important problem class has received little attention in the literature, since the convex component of the objective does not enjoy strong convexity.
Read the full abstract

All authors Andrew Calcan, Jordan Collard, Alberto De Marchi, Scott B. Lindstrom

Optimization and Control

Robust Optimal Reinsurance, Investment,and Surplus Allocation for Epstein-Zin Preferences

In this paper, we investigate the robust optimal reinsurance,investment,and internal surplus distribution (i.e., consumption) problem for an insurer with Epstein-Zin recursive preferences in an incomplete market. It is assumed that the insurer can allocate wealth to a financial market consisting of a risk-free asset and a risky asset, where the price process of the risky asset follows a diffusion process with a stochastic drift rate governed by an Ornstein-Uhlenbeck (O-U) process. For both the unit and non-unit elasticity of intertemporal substitution (EIS) cases, by applying the classical dynamic programming approach, we derive explicit solutions for the optimal robust reinsurance, investment,and consumption strategies and also verify that the obtained solutions indeed solve the optimal control problem. Furthermore, we compare the robust solutions with their non-robust counterparts, and the comparative results shown in the figures are consistent with economic intuition. Finally, we contrast the exact solutions with the Campbell-Shiller approximation and assess the accuracy of the approximation method.
Read the full abstract

Probability

Lévy processes with partially stochastic resetting

In this paper, we solve exit problems for a Lévy process that resets proportionally to its current position at independent Poisson epochs times. This resetting causes an additional (proportional to its current level) downward (upward) jump when the current position of the process is on the positive (negative) domain. Such a process can be expressed as an SDE, whose existence and uniqueness it discussed. All identities are given in terms of new family of scale functions. To obtain the new family of scale functions, we reduce the problem of the LT of the exit times into integral equations that are solve in terms of resolvent series.
Read the full abstract

All authors Zbigniew Palmowski, Noah Beelders, Lewis Ramsden, Apostolos D. Papaioannou

Analysis of PDEs

Stability of Anomalous Dissipation for the Forced 3D Navier–Stokes Equations under Geometric Perturbations

The energy dissipation in the inviscid limit is a central problem in turbulence theory. Kolmogorov's K41 theory predicts a positive dissipation rate independent of viscosity – a phenomenon known as anomalous dissipation. Brué and De Lellis gave the first rigorous construction, but it relies on extremely precise geometric conditions. Based on quasi-self-similar mixing, we prove structural stability under pure normal perturbations of the central curves. We establish C^2 stability of the maps and C^1 stability of the local fields, and obtain Hölder estimates and high-frequency energy concentration. A contradiction gives a positive dissipation lower bound independent of the perturbation, and embedding into the (2+1/2)-dimensional framework shows C^6 structural stability. The main novelty is that the Brué–De Lellis construction remains stable under such perturbations, so anomalous dissipation occurs in an open neighbourhood of function spaces, providing a rigorous foundation for K41 theory.
Read the full abstract

Logic

Fractional Helly property and combinatorics of forking in NTP$_2$ theories

We investigate the class of FHP theories, i.e. theories of structures in which all definable families of sets satisfy the Fractional Helly Property (and its variants) from combinatorics. FHP theories generalize NIP and form a new subclass of low NTP$_2$ theories. We give many new examples (including ultraproducts of finite fields and of the $p$-adics) and establish some results about forking and $f$-generics for amenable groups definable in FHP theories. We make several conjectures about finitary combinatorial properties of forking in NTP$_2$ theories and establish some partial results, as well as investigate related two-cardinal type counting functions addressing a question of Adler.
Read the full abstract

Dynamical Systems

Existence of Solutions for Hybrid Dynamical Systems on Graph State-Space

This paper proposes a framework to ensure the existence of dynamical system trajectories in the state space of labeled, weighted, and attributed graphs. The evolution of such a system exhibits hybrid behavior: discrete jumps affecting the topology-the emergence and disappearance of vertices and edges over time-as well as vertex attributes and edge weights, combined with a continuous evolution of these same attributes and weights. To address the discrete behavior, an appropriate algebraic structure for the graph space is proposed; the analysis of its properties shows the existence of a new mathematical structure: a semi-vector space over the field of real numbers, whereas the literature only describes semi-vector spaces over semi-fields. The continuous behavior is modeled by differential equations. To facilitate formal treatment, the graph space is embedded, via a semi-linear mapping, into a new space, called variable-basis space, introduced in this work. The system's evolution model is then formulated within this space-the image of the graph space in the variable-basis space-under the framework of hybrid dynamical systems theory. A general result on the existence of solutions is established. Finally, this framework is applied to model and simulate the dynamics of the gut microbiota under antibiotic treatment followed by bacteriotherapy, where a generalized Lotka-Volterra (gLV) model describes the evolution of species abundances.
Read the full abstract

All authors Arthur Doliveira, Christophe Roman, Guillaume Graton, Mustapha Ouladsine

Rings and Algebras

Weakly Noetherian Lie Algebra and the Sierra-Walton Conjecture

Let K be a field of characteristic zero. Motivated by the conjecture that an enveloping algebra U(g) is Noetherian only if g is finite dimensional, we define the notion of weakly Noetherian Lie algebras. The main result, Theorem A, states that weakly Noetherian Lie algebras have a very constrained structure. In the specific case of graded Lie algebras, it implies an explicit classification of the perfect strictly weakly Noetherian Lie algebras, stated in Theorem B. The proofs of both theorems are quite long, and uses concrete results due to Tits, Formanek, Razmyslov, Grabowski and the author. The first theorem provides some insight on the desired conjecture. The second one implies the conjecture for all perfect graded Lie algebras, improving a celebrated theorem of Sierra and Walton.
Read the full abstract

Dynamical Systems

Singularity Collisions through Homotopical Dynamical cancellation

We introduce collisions of invariant sets and, in particular, consider dynamical homotopical cancellations that preserve the homotopy type of the underlying singular manifold. We develop the theory of homotopical dynamical cancellation for generalized Gutierrez-Sotomayor (GGS) flows defined on GGS manifolds. This framework extends the classical cancellation theory of Morse flows to the singular setting. To effectively capture these homotopical cancellations, we introduce a GGS chain complex, which encodes essential dynamical and algebraic-topological information. Furthermore, we provide a spectral sequence analysis of a filtered GGS chain complex, demonstrating a bijective correspondence between algebraic cancellations of the modules of the spectral sequence and homotopical dynamical cancellations in the GGS flow. Several illustrative examples are presented, highlighting the practical applicability of the proposed framework.
Read the full abstract

Algebraic Geometry · Cross-listed from cs.SC

Computing points in connected components defined by a real inequation: algorithms, complexity and implementations, Part I

We consider the problem of computing sample points in each connected component of a semi-algebraic set defined by the non-vanishing or the positivity of an n-variate polynomial of degree d, with rational coefficients of bit size bounded by $\tau$. Such a problem is a basic routine in effective real algebraic geometry, used in higher-level algorithms for solving polynomial systems over the reals and finds many applications in sciences. We design a probabilistic algorithm for solving this problem, which is based on reductions to different routines for solving zero-dimensional polynomial systems. It assumes that the input polynomial satisfies sufficiently generic properties (namely, smoothness of its defining hypersurface). This is done through the computations of critical points of well-chosen maps to capture the connected components of the semi-algebraic set under study. We derive a bit complexity estimate for the cost of this algorithm, which is, in terms of the Bézout bound d(d -1)^{n-1}, essentially cubic for obtaining parametrisations of the sought-for real points. Moreover, we also consider the case of obtaining rational approximations of those points, which are precise enough to lie in the same connected components as their exact counterparts, which yields a cost that is essentially quartic in the Bézout bound. In these complexity estimates, we take into account the degree structure of the input polynomial and its partial derivatives, allowing for a more refined bit complexity when the partial derivative of the input polynomial have degree lower than expected. We also analyse the probability of success of those algorithms. We report on practical experiments, benchmarking with random dense input polynomials as well as polynomials coming from applications, which were out of reach of the state-of-the-art implementations, and hence illustrate the practical efficiency of these new algorithms.
Read the full abstract

Classical Analysis and ODEs

What lies between polynomial and exponential growth?

In this paper we give an alternative exposition of a recent paper regarding the classification of growth rates of real functions. We take a different point of view, focussing on understanding possible growth rates between polynomial and exponential. In order to be able to explicitly name a range of such functions, we first need to extend our basic functions. We do this via a 'tower' of Abel functions. With these one can classify functions in a natural way with polynomials and exponentials in consecutive classes. We show there are large gaps between these classes which indicate that it is mostly unknown what lies between polynomial and exponential growth, especially if the "Continuum Hypothesis for classes" is true.
Read the full abstract

Optimization and Control

Symmetry-Compatible Principle for Optimizer Design: Embeddings, LM Heads, SwiGLU MLPs, and MoE Routers

A striking geometric disparity has long persisted in the practice of deep learning. While modern neural network architectures naturally exhibit rich symmetry and equivariance properties, popular optimizers such as Adam and its variants operate inherently coordinate-wise, rendering them unable to respect the equivariance structures of the parameter space. We address this disparity by introducing a symmetry-compatible principle for optimizer design: the gradient update rule should be equivariant under the symmetry group acting on the corresponding weight block. Following this principle, we first provide a unified perspective on bi-orthogonally equivariant updates for general matrix layers, as employed by stochastic spectral descent, Muon, Scion, and polar gradient methods. More importantly, by moving from orthogonal groups to permutation and shared-shift symmetries, we derive symmetry-compatible optimizers for parameter blocks whose symmetries differ from those of general matrix layers: embedding and LM head matrices, SwiGLU MLP projections, and MoE router matrices. These constructions include one-sided spectral, row-norm, hybrid row-norm/spectral, row-aware, column-aware, centered row-norm, and left-spectral updates. They yield an end-to-end layerwise optimizer stack in which each major matrix-valued parameter class is assigned an update whose equivariance matches its symmetry group. We corroborate this principle through pre-training experiments on dense and sparse MoE language models, including Qwen3-0.6B-style, Gemma 3 1B-style, OLMoE-1B-7B-style, and downsized gpt-oss architectures. Across these experiments, symmetry-compatible update rules consistently improve final validation loss, reduce expert load imbalance in sparse MoE models, and in several cases control final vocabulary-logit growth, improve router stability, and overall training stability over the corresponding AdamW updates.
Read the full abstract

Functional Analysis

Positive Linear Maps on Second Symmetric Product Spaces

Let $X^{(2)}$ denote the second symmetric product space of a partially ordered vector space $X$, endowed with the projective cone. A characterization of linear maps $T\colon X^{(2)}\to X^{(2)}$ which preserve the set of all positive decomposable vectors, is proved. As applications of this result, an alternative proof, as well as an infinite dimensional generalization, of a representation theorem for (i) automorphisms on the completely positive cone and (ii) linear preservers of CP-rank-1 matrices, are presented. It is also shown that if $T$ preserves the set of all decomposable vectors, then so does the Drazin inverse, $T^D$ (if it exists). The case of the Moore-Penrose inverse is also investigated.
Read the full abstract

Statistics Theory

Uncertainty functionals revisited: Concavity and Jensen's inequality

This article presents a theoretical study of uncertainty functionals on general measurable spaces. These functionals are fundamental in experimental design and global sensitivity analysis, where they are used to quantify variability and information content in probabilistic models. As first articulated in DeGroot's seminal 1962 article, a natural requirement is that uncertainty should decrease on average when additional information is obtained. This requirement is equivalent to the probabilistic form of Jensen's inequality on the space of probability measures. Our main results show that concavity is necessary but not sufficient for Jensen's inequality to hold whenever the underlying measurable space is infinite. We also provide practicable sufficient conditions under which the desired property holds. These results contribute to a clearer mathematical foundation for uncertainty quantification. Several open questions are formulated.
Read the full abstract

Information Theory

Fluid Antenna-Aided Robust Secure Uplink Transmission in LEO Satellite Constellations

Secure uplink transmission in low-earth orbit (LEO) satellite constellations faces significant challenges due to the time-varying spatial distributions of serving and potential eavesdropping satellites, particularly when their angular separations are small and eavesdropper channel state information (CSI) is imperfect. To address these challenges, this paper investigates fluid antenna (FA)-aided robust secure uplink transmission in LEO satellite constellations. To avoid frequent antenna reconfiguration, the FA positions are optimized over an observation interval and remain fixed throughout the interval, while transmit beamforming dynamically adapts to the time-varying satellite channels. We formulate an average worst-case secrecy rate maximization problem by jointly optimizing FA positions and time-varying beamforming under norm-bounded eavesdropper CSI uncertainties and per-slot secrecy rate constraints. Two new alternating optimization (AO) algorithms are developed to solve the resulting non-convex problem. For robust beamforming optimization, the S-procedure, semidefinite relaxation (SDR), and first-order approximation are employed to construct a tractable convex problem, followed by Gaussian randomization for rank-one recovery. For FA position optimization, we derive concave surrogate functions for successive convex approximation (SCA) and develop a differential evolution (DE)-based search strategy to reduce computational complexity. Numerical results demonstrate that both FA-aided schemes consistently outperform conventional fixed-position antenna arrays across different constellation configurations and CSI uncertainty levels. Moreover, the two algorithms exhibit different performance-complexity tradeoffs as the number of FA elements increases, with DE offering a computational advantage for larger arrays while SCA can achieve higher secrecy rates.
Read the full abstract

All authors Suhong Luo, Pan Tang, Jianhua Zhang, Ji Wang, Yixuan Li, Xingwang Li

Numerical Analysis

A regularization method for planar offset curves and bi-offset recognition

Offset curves for planar trajectories are interesting in the generation of tool paths for numerically controlled industrial machines and in trajectory planning methods for autonomous driving systems. Theoretical offset curves may exhibit peculiar singularities, including self-intersections, which limit their use in practical applications. Existing approaches address these issue through geometric filtering techniques to detect and remove undesirable features but the computation of accurate and well-behaved offset curves remains a challenging task. We assume a first stage of functional approximation of trajectories by penalized Hermite spline regression enabling the simultaneous fitting of positions and tangents. The regularization is imposed on the second derivatives, effectively mitigating the jerk effect, which is particularly relevant in motion planning and path smoothing applications. Then, taking into account the geometrical pointwise properties of the resulting curve, we design two offset curves through the simultaneous approximation of function values and derivatives. Then, a mathematical model to obtain the so-called bi-offset as most fitting as with the original generator curve is proposed, also relating the offset range and pointwise curvature values. The adaptive reconstruction of the center line from the external boundaries is a topic of interest and is the main focus of our work. Numerical experiments confirm the reliability of our approach at every stage of the resolution process.
Read the full abstract

Mathematical Physics

Where solitons are in a KdV soliton gas

The Korteweg-De Vries (KdV) equation is a paradigmatic model of integrable classical fields, with solitonic solutions where localised profiles interact by elastic, factorised scattering. When solitons interact, their shapes are modified, and individual solitons cannot be identified. This is a key problem in soliton gas where one needs a density of solitons per unit distance: can we locate asymptotic objects, quasiparticles, even at finite times in the macroscopic limit? Quasiparticles' positions should satisfy locality properties, the field in mesoscopic regions being determined by quasiparticles lying there. We solve this problem for the KdV equation. We define solitons' positions as functions of spectral and impact parameters via an expansion bijection generalising that of hard rods. These “KdV rods” solve the semiclassical Bethe equations, implementing the collision rate ansatz, and satisfy a repulsion condition generalising the rods' lengths to the KdV scattering shifts and related to the condensate condition. This gives an explicit characterisation of multi-soliton fields in terms of solitons' positions instead of impact parameters. We construct fluid-cell projections, that project out solitons lying outside mesoscopic regions: the resulting field is unchanged in this region and supported there up to small corrections. We show that in the weak limit, conserved densities are given by the empirical density for these solitons' positions. Thus we have associated a kinetic theory of quasiparticles to the KdV equation. The results hold under natural assumptions on spectral and impact parameters, and no randomness is required. Our proof is based on novel solitonic tau functions. This is a crucial step towards rigorously deriving the kinetic equation of the KdV soliton gas, first proposed by Gennady El in 2003. We believe the methods are generalisable to other solitonic models.
Read the full abstract

2605.18093 · 18 May · v1: 85 pp, 3 figs. v2: 109 pp, 3 figs. Exact solution to SBE. Significantly simplified assumptions, clear physical meaning, substantially strengthen results. Analysis based on SBE instead of magnifying-glass positions. v3: 113 pp, 4 figs, typos corrected, assumptions further simplified."KdV rods" with clear separation conditions, proof that SBE is bijection between this and impact parameters

Algebraic Geometry

Chern classes of Laughlin bundles on the quasihole moduli space

We study fractional quantum Hall states with quasihole excitations, on Riemann surfaces of arbitrary genus. For configurations with $m$ quasiholes we construct a vector bundle above the $m$-th symmetric power of the curve so that the fiber at a point $\lbrace w_1,\dots,w_m \rbrace$ corresponds to the state with quasiholes localized at these positions. We determine the Chern character of this bundle via the Grothendieck-Riemann-Roch theorem and show that in the completely filled state, i.e. when the number of particles is maximal, the vector bundle is compatible with the condition of projective flatness. Furthermore, we obtain a generalization of this result to the case of multiple layers and multiple quasihole types. In genus zero and one, we construct explicit wave-functions and verify that the curvature of the associated Chern connection reproduces the predicted Chern classes. The Chern classes obtained match, term by term, the predicted decomposition of the Berry phase under quasihole exchange, into an extensive Aharonov–Bohm contribution and a fractional statistical contribution.
Read the full abstract

Category Theory

Enriched categories, real metrics and Lorentz manifolds

This expository article brings together two subjects: generalised metrics based on enriched categories, on the one hand, and Lorentz manifolds, on the other, at the price of dealing with details that are well known either in category theory or in relativity. The spacetime of relativity can be given a real valued metric $\rho(x, y)$, with values in the extended real line, or better (if equivalently) a real valued `antimetric' $\gamma(x, y) = - \rho(x, y)$ (satisfying a reverse triangle inequality); the latter, as a function of $y$, is positive on the timecone of $x$, annihilates on its lightcone, and is $- \infty$ on all events which cannot be influenced by $x$. All this can be given a well-established base in category theory, extending Lawvere's notion of a metric space. In fact, a space with a real valued metric can be viewed as an enriched category on the extended real line, structured as a symmetric monoidal closed category.
Read the full abstract

Classical Analysis and ODEs

Sharp Estimates for Conjugate Functions with Applications to Trigonometric Polynomials

We prove a sharp estimate for conjugate functions using a harmonic majorant in a half-strip. As an application, we remove the logarithmic loss from a theorem of Papadopoulos on minima of trigonometric polynomials and obtain the optimal order in the corresponding inequality.

Information Theory

A Hexagonal Counterexample to Log-Convexity of Fisher Information Along the Heat Flow

We construct a smooth, strictly positive, Gaussian-decaying density on $\mathbb{R}^2$ for which Fisher information along the heat flow is not log-convex. This disproves the Cheng–Geng log-convexity conjecture in dimension two and, by tensorization, in every dimension $d\ge2$. Consequently, the multidimensional forms of the Gaussian completely monotone conjecture, McKean's conjecture, and Toscani's entropy power conjecture also fail, complementing the one-dimensional counterexample of Gu and Sellke. Our construction is a small hexagonal perturbation on the triangular torus, transferred to $\mathbb{R}^2$ by a Gaussian envelope and supported by explicit two-dimensional numerics. We also initiate the study of the sharp constants $\theta_d^*$ by proving $\theta_1^*=1$, establishing monotonicity in the dimension, and identifying a dichotomy for the asymptotic constant $\theta_\infty^*$ governed by the sign of $\mathcal{D}$. The explicit two-dimensional counterexample was found by GPT-5.5 Pro.
Read the full abstract

All authors Jiayang Zou, Luyao Fan, Jiayang Gao, Jia Wang

Probability · Cross-listed from stat.ML

Wasserstein bounds for denoising diffusion probabilistic models via the Föllmer process

This paper studies sampling error bounds for denoising diffusion probabilistic models (DDPMs) in the 2-Wasserstein distance. Our contributions are threefold. (i) Under general Lipschitz-type conditions on the score function and for a broad class of variance schedules, including the cosine schedule, we establish sharp upper bounds that are optimal in both the dimension and the number of steps, and recover several sharp error bounds previously obtained in the literature. (ii) We prove that the same Lipschitz-type conditions, which encompass those commonly imposed on the (learned) score, imply a logarithmic Sobolev inequality and hence a quadratic transportation cost inequality for the DDPM. As a consequence, in settings covered by existing work, an optimal Wasserstein bound, up to a logarithmic factor, follows from the recently obtained sharp error bound in the Kullback-Leibler divergence under geometric-type variance schedules. (iii) We show that for general log-concave target distributions, the optimal Wasserstein error bound remains attainable even without a quadratic transportation cost inequality for the target. Our analysis is based on viewing the DDPM sampler as a discretization of the Föllmer process rather than the conventional reverse Ornstein-Uhlenbeck process.
Read the full abstract

Algebraic Geometry

Limits of Hyperkähler Manifolds and Moduli Spaces of Sheaves on K3 Surfaces

We prove that a compact strongly Gauduchon manifold which is a limit of compact hyperkähler manifolds is bimeromorphic to a compact hyperkähler manifold. This settles a case of a conjecture of Soldatenkov and Verbitsky. A closed completion of a Gauduchon metric bounds the Dirichlet energies of nearby biholomorphisms. This bound yields a meromorphic limit map, which has degree one and is therefore a bimeromorphism onto a compact hyperkähler manifold. We apply this result to moduli spaces of stable torsion-free coherent sheaves on K3 surfaces. A Bott–Chern diagonal formula gives the $\partial\bar\partial$-lemma, and positivity of determinant classes gives Kählerness, proving Perego's conjecture.
Read the full abstract

History and Overview of Mathematics

Meeting Solomon Marcus

Dedicated to Solomon Marcus, the current paper continues a recent series about our meetings. Trying to recreate the spirit of those meetings, we first propose a discussion which started as a high-school problem. The main part of the current paper consists in a section about racks. It presents elements of trigonometry in racks, and Euler formulas associated in this framework
Read the full abstract

Analysis of PDEs

The omnidirectional trace in H 1 ($\Omega$)

We first prove that all the functions in L 2 whose directional derivative is in L 2 have a directional trace on the boundary of any open bounded domain, without assumptions on its regularity. This enables us to define the omnidirectional trace of the elements of the Sobolev space H 1 ($\Omega$) for which there exists a function on the boundary that is almost everywhere equal, with respect to the directional measure, to the directional trace, regardless of the direction. The set of all these elements of H 1 ($\Omega$), denoted by H 1 tr ($\Omega$), is shown to be closed, and to always contain the closure in H 1 ($\Omega$) of the set C 0 ($\Omega$)$\cap$ H 1 ($\Omega$) (it is always equal to this set in the 1D case, and can be strictly greater in higher dimensions). The omnidirectional trace always satisfies an integration-by-parts formula, which combines the values of the trace on opposite points of the boundary. Examples show that this notion enables the resolution of variational problems involving the values at the boundary of the domain.
Read the full abstract

All authors Robert Eymard, Thierry Gallouët, David Maltese, Lucas Oger

Probability · Cross-listed from stat.ML

Connections between the Föllmer process and the denoising diffusion probabilistic model

The Föllmer process is a Brownian motion conditioned to have a pre-specified distribution at time 1. This process can be interpreted as an “augmented” time-compressed version of the reverse stochastic differential equation (SDE) corresponding to the denoising diffusion probabilistic model (DDPM). While this fact has been indirectly used to analyze DDPM sampling errors via discretization of the reverse SDE, the connection between direct discretization of the Föllmer process and the DDPM sampler has not yet been fully explored. This paper clarifies this point while surveying relevant results from the literature. We show that discretized Föllmer processes give natural hyper-parameter settings of the DDPM sampler while accommodating a broader class of variance schedules than discretized reverse SDEs. Moreover, this allows us to systematically recover state-of-the-art results on DDPM sampling error bounds, along with slight improvements.
Read the full abstract

Combinatorics

On Occurrence-Preserving Morphisms

A morphism is a mapping that transforms words through letter-wise substitution, where each symbol is consistently replaced by a fixed word. In the field of combinatorics on words, one topic that has attracted considerable attention is the characterization of morphisms that preserve specific properties, such as overlap-freeness, square-freeness, lexicographic order, and primitivity. Continuing this direction, we initiate the study on occurrence-preserving morphisms, which address the following fundamental question: given a morphism $\phi$, two words $u$ and $v$, and $k \geq 1$, under what conditions does the number of occurrences of $u$ in $v$ equal the number of occurrences of $\phi^k(u)$ in $\phi^k(v)$? To answer this question, we introduce the notion of interference-free morphisms, examine their properties, develop an efficient algorithm for deciding interference-freeness, and uncover a connection to recognizable morphisms. We then present a precise characterization of occurrence-preserving morphisms in terms of interference-freeness. As applications of our characterization, we first show that there exists a bijection between the starting positions of the occurrences of $u$ in $v$ and those of $\phi^k(u)$ in $\phi^k(v)$. We then apply the characterization to the Fibonacci and Thue-Morse words to identify their minimal unique substrings~(MUSs). Finally, we exploit the connection between MUSs and net occurrences to simplify existing proofs on net occurrences in these words.
Read the full abstract

All authors Kaisei Kishi, Peaker Guo, Cristian Urbina, Hideo Bannai